chore: release v3.27.2

* add tests

* push

.env.dev.legacy.example

@@ -1,86 +0,0 @@
-# When adding additional environment variables, the schema in "/src/env.mjs"
-# should be updated accordingly.
-
-# Prisma
-# https://www.prisma.io/docs/reference/database-reference/connection-urls#env
-DIRECT_URL="postgresql://postgres:postgres@localhost:5432/postgres"
-DATABASE_URL="postgresql://postgres:postgres@localhost:5432/postgres"
-
-# Clickhouse
-CLICKHOUSE_MIGRATION_URL="clickhouse://localhost:9000"
-CLICKHOUSE_URL="http://localhost:8123"
-CLICKHOUSE_USER="clickhouse"
-CLICKHOUSE_PASSWORD="clickhouse"
-CLICKHOUSE_CLUSTER_ENABLED="false"
-
-# Next Auth
-# You can generate a new secret on the command line with:
-# openssl rand -base64 32
-# https://next-auth.js.org/configuration/options#secret
-# NEXTAUTH_SECRET=""
-NEXTAUTH_URL="http://localhost:3000"
-NEXTAUTH_SECRET="secret"
-
-# Langfuse Cloud Environment
-NEXT_PUBLIC_LANGFUSE_CLOUD_REGION="DEV"
-
-# Langfuse experimental features
-LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES="true"
-
-# Salt for API key hashing
-SALT="salt"
-
-# Email
-EMAIL_FROM_ADDRESS=""  # Defines the email address to use as the from address.
-SMTP_CONNECTION_URL=""  # Defines the connection url for smtp server.
-
-# S3 Batch Exports
-LANGFUSE_S3_BATCH_EXPORT_ENABLED=true
-LANGFUSE_S3_BATCH_EXPORT_BUCKET=langfuse
-LANGFUSE_S3_BATCH_EXPORT_ACCESS_KEY_ID=minio
-LANGFUSE_S3_BATCH_EXPORT_SECRET_ACCESS_KEY=miniosecret
-LANGFUSE_S3_BATCH_EXPORT_REGION=us-east-1
-LANGFUSE_S3_BATCH_EXPORT_ENDPOINT=http://localhost:9090
-## Necessary for minio compatibility
-LANGFUSE_S3_BATCH_EXPORT_FORCE_PATH_STYLE=true
-LANGFUSE_S3_BATCH_EXPORT_PREFIX=exports/
-
-# # S3 Media Upload LOCAL
-LANGFUSE_S3_MEDIA_UPLOAD_BUCKET=langfuse
-LANGFUSE_S3_MEDIA_UPLOAD_ACCESS_KEY_ID=minio
-LANGFUSE_S3_MEDIA_UPLOAD_SECRET_ACCESS_KEY=miniosecret
-LANGFUSE_S3_MEDIA_UPLOAD_REGION=us-east-1
-LANGFUSE_S3_MEDIA_UPLOAD_ENDPOINT=http://localhost:9090
-## Necessary for minio compatibility
-LANGFUSE_S3_MEDIA_UPLOAD_FORCE_PATH_STYLE=true
-LANGFUSE_S3_MEDIA_UPLOAD_PREFIX=media/
-
-# S3 Event Bucket Upload
-## Set to true to test uploading all events to S3
-LANGFUSE_S3_EVENT_UPLOAD_BUCKET=langfuse
-LANGFUSE_S3_EVENT_UPLOAD_ACCESS_KEY_ID=minio
-LANGFUSE_S3_EVENT_UPLOAD_SECRET_ACCESS_KEY=miniosecret
-LANGFUSE_S3_EVENT_UPLOAD_REGION=us-east-1
-LANGFUSE_S3_EVENT_UPLOAD_ENDPOINT=http://localhost:9090
-## Necessary for minio compatibility
-LANGFUSE_S3_EVENT_UPLOAD_FORCE_PATH_STYLE=true
-LANGFUSE_S3_EVENT_UPLOAD_PREFIX=events/
-
-# Set during docker build of application
-# Used to disable environment verification at build time
-# DOCKER_BUILD=1
-
-REDIS_HOST="127.0.0.1"
-REDIS_PORT=6379
-REDIS_AUTH="myredissecret"
-
-# openssl rand -hex 32 used only here
-ENCRYPTION_KEY=0000000000000000000000000000000000000000000000000000000000000000 
-
-# speeds up local development by not executing init scripts on server startup
-NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT="false"
-
-LANGFUSE_READ_FROM_POSTGRES_ONLY=true
-LANGFUSE_RETURN_FROM_CLICKHOUSE=false
-LANGFUSE_READ_DASHBOARDS_FROM_CLICKHOUSE=false
-LANGFUSE_READ_FROM_CLICKHOUSE_ONLY=false

.env.prod.example

@@ -55,6 +55,7 @@ OTEL_SERVICE_NAME="langfuse"
 
 # Auth, optional configuration
 # AUTH_DOMAINS_WITH_SSO_ENFORCEMENT=domain1.com,domain2.com
+# AUTH_IGNORE_ACCOUNT_FIELDS=foo,bar
 # AUTH_DISABLE_USERNAME_PASSWORD=true
 # AUTH_DISABLE_SIGNUP=true
 # AUTH_SESSION_MAX_AGE=43200 # 30 days in minutes (default)
@@ -64,43 +65,65 @@ OTEL_SERVICE_NAME="langfuse"
 # AUTH_GOOGLE_CLIENT_SECRET=
 # AUTH_GOOGLE_ALLOW_ACCOUNT_LINKING=false
 # AUTH_GOOGLE_ALLOWED_DOMAINS=langfuse.com,google.com # optional allowlist of workspace domains that can sign in via Google
+# AUTH_GOOGLE_CLIENT_AUTH_METHOD=
+# AUTH_GOOGLE_CHECKS=
 # AUTH_GITHUB_CLIENT_ID=
 # AUTH_GITHUB_CLIENT_SECRET=
 # AUTH_GITHUB_ALLOW_ACCOUNT_LINKING=false
+# AUTH_GITHUB_CLIENT_AUTH_METHOD=
+# AUTH_GITHUB_CHECKS=
 # AUTH_GITHUB_ENTERPRISE_CLIENT_ID=
 # AUTH_GITHUB_ENTERPRISE_CLIENT_SECRET=
 # AUTH_GITHUB_ENTERPRISE_BASE_URL=
 # AUTH_GITHUB_ENTERPRISE_ALLOW_ACCOUNT_LINKING=false
+# AUTH_GITHUB_ENTERPRISE_CLIENT_AUTH_METHOD=
+# AUTH_GITHUB_ENTERPRISE_CHECKS=
 # AUTH_GITLAB_CLIENT_ID=
 # AUTH_GITLAB_CLIENT_SECRET=
 # AUTH_GITLAB_ALLOW_ACCOUNT_LINKING=false
 # AUTH_GITLAB_ISSUER=
+# AUTH_GITLAB_CLIENT_AUTH_METHOD=
+# AUTH_GITLAB_CHECKS=
 # AUTH_AZURE_AD_CLIENT_ID=
 # AUTH_AZURE_AD_CLIENT_SECRET=
 # AUTH_AZURE_AD_TENANT_ID=
-# AUTH_AZURE_ALLOW_ACCOUNT_LINKING=false
+# AUTH_AZURE_AD_ALLOW_ACCOUNT_LINKING=false
+# AUTH_AZURE_AD_CLIENT_AUTH_METHOD=
+# AUTH_AZURE_AD_CHECKS=
 # AUTH_OKTA_CLIENT_ID=
 # AUTH_OKTA_CLIENT_SECRET=
 # AUTH_OKTA_ISSUER=
 # AUTH_OKTA_ALLOW_ACCOUNT_LINKING=false
+# AUTH_OKTA_CLIENT_AUTH_METHOD=
+# AUTH_OKTA_CHECKS=
 # AUTH_AUTH0_CLIENT_ID=
 # AUTH_AUTH0_CLIENT_SECRET=
 # AUTH_AUTH0_ISSUER=
 # AUTH_AUTH0_ALLOW_ACCOUNT_LINKING=false
+# AUTH_AUTH0_CLIENT_AUTH_METHOD=
+# AUTH_AUTH0_CHECKS=
 # AUTH_COGNITO_CLIENT_ID=
 # AUTH_COGNITO_CLIENT_SECRET=
 # AUTH_COGNITO_ISSUER=
 # AUTH_COGNITO_ALLOW_ACCOUNT_LINKING=false
+# AUTH_COGNITO_CLIENT_AUTH_METHOD=
+# AUTH_COGNITO_CHECKS=
 # AUTH_KEYCLOAK_CLIENT_ID=
 # AUTH_KEYCLOAK_CLIENT_SECRET=
 # AUTH_KEYCLOAK_ISSUER=
 # AUTH_KEYCLOAK_ALLOW_ACCOUNT_LINKING=false
+# AUTH_KEYCLOAK_CLIENT_AUTH_METHOD=
+# AUTH_KEYCLOAK_CHECKS=
 # AUTH_CUSTOM_CLIENT_ID=
 # AUTH_CUSTOM_CLIENT_SECRET=
 # AUTH_CUSTOM_ISSUER=
 # AUTH_CUSTOM_NAME=
 # AUTH_CUSTOM_SCOPE="openid email profile" # optional
+# AUTH_CUSTOM_CLIENT_AUTH_METHOD="client_secret_basic" # optional
 # AUTH_CUSTOM_ALLOW_ACCOUNT_LINKING=false
+# AUTH_CUSTOM_ID_TOKEN=false # optional, default is true
+# AUTH_CUSTOM_CLIENT_AUTH_METHOD=
+# AUTH_CUSTOM_CHECKS=
 
 # Transactional email, optional
 # Defines the email address to use as the from address.
@@ -243,13 +266,12 @@ OTEL_SERVICE_NAME="langfuse"
 # CLICKHOUSE_URL=
 # CLICKHOUSE_USER=
 # CLICKHOUSE_PASSWORD=
+# CLICKHOUSE_DB=
 
 # Ingestion
 # LANGFUSE_INGESTION_QUEUE_DELAY_MS=
 # LANGFUSE_INGESTION_CLICKHOUSE_WRITE_BATCH_SIZE=
 # LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS= 
 # LANGFUSE_INGESTION_CLICKHOUSE_MAX_ATTEMPTS=
-# LANGFUSE_LEGACY_INGESTION_WORKER_CONCURRENCY=
-# QUEUE_CONSUMER_LEGACY_INGESTION_QUEUE_IS_ENABLED="true"
 
-## END Langfuse V3 Ingestion
+## END Langfuse V3 Ingestion

.github/workflows/codespell.yml

@@ -3,9 +3,13 @@ name: Codespell
 
 on:
   push:
-    branches: [main]
+    branches:
+      - "main"
+    tags:
+      - "v*"
   pull_request:
-    branches: [main]
+    branches:
+      - "**"
   merge_group:
 
 permissions:

.github/workflows/pipeline.yml

@@ -17,8 +17,22 @@ concurrency:
   cancel-in-progress: ${{ github.event_name == 'pull_request' }}
 
 jobs:
+  pre-job:
+    runs-on: ubuntu-latest
+    outputs:
+      should_skip: ${{ steps.skip_check.outputs.should_skip }}
+    timeout-minutes: 15
+    steps:
+      - id: skip_check
+        uses: fkirc/skip-duplicate-actions@v5
+        with:
+          do_not_skip: '["workflow_dispatch"]'
+
   lint:
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     steps:
       - uses: actions/checkout@v4
       - uses: pnpm/action-setup@v3
@@ -41,6 +55,9 @@ jobs:
   test-docker-build:
     timeout-minutes: 20
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     steps:
       - name: Checkout
         uses: actions/checkout@v4
@@ -74,6 +91,9 @@ jobs:
   tests-web-sync:
     timeout-minutes: 20
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     name: tests-web-sync (node${{ matrix.node-version }}, pg${{ matrix.postgres-version }})
     strategy:
       matrix:
@@ -87,7 +107,7 @@ jobs:
       - uses: actions/checkout@v4
       - name: Install golang-migrate for Clickhouse migrations
         run: |
-          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.16.2/migrate.linux-amd64.tar.gz | tar xvz
+          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.18.2/migrate.linux-amd64.tar.gz | tar xvz
           sudo mv migrate /usr/bin/migrate
           which migrate
       - uses: pnpm/action-setup@v3
@@ -111,8 +131,9 @@ jobs:
       - name: Load default env
         run: |
           cp .env.dev.example .env
-          grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^REDIS_HOST=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev.legacy.example > .env
-          echo "LANGFUSE_POSTGRES_INGESTION_ENABLED=true" >> .env
+          grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev.example > .env
+          echo "LANGFUSE_INGESTION_QUEUE_DELAY_MS=1" >> .env
+          echo "LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=1" >> .env
       - name: Run + migrate
         run: |
           docker compose -f docker-compose.dev.yml up -d
@@ -146,6 +167,9 @@ jobs:
   tests-web-async:
     timeout-minutes: 20
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     name: tests-web-async (node${{ matrix.node-version }}, pg${{ matrix.postgres-version }}, mode${{ matrix.blob-provider }})
     strategy:
       matrix:
@@ -160,7 +184,7 @@ jobs:
       - uses: actions/checkout@v4
       - name: Install golang-migrate for Clickhouse migrations
         run: |
-          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.16.2/migrate.linux-amd64.tar.gz | tar xvz
+          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.18.2/migrate.linux-amd64.tar.gz | tar xvz
           sudo mv migrate /usr/bin/migrate
           which migrate
       - uses: pnpm/action-setup@v3
@@ -184,7 +208,9 @@ jobs:
       - name: Load default env
         run: |
           cp .env.dev${{ matrix.blob-provider }}.example .env
-          grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^REDIS_HOST=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev${{ matrix.blob-provider }}.example > .env
+          grep -v -e '^LANGFUSE_S3_BATCH_EXPORT_ENABLED=' -e '^NEXT_PUBLIC_LANGFUSE_RUN_NEXT_INIT=' .env.dev${{ matrix.blob-provider }}.example > .env
+          echo "LANGFUSE_INGESTION_QUEUE_DELAY_MS=1" >> .env
+          echo "LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS=1" >> .env
       - name: Run + migrate
         run: |
           docker compose -f docker-compose.dev${{ matrix.blob-provider }}.yml up -d
@@ -216,6 +242,9 @@ jobs:
   tests-worker:
     timeout-minutes: 20
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     name: tests-worker (node${{ matrix.node-version }}, pg${{ matrix.postgres-version }}, mode${{ matrix.blob-provider }})
     strategy:
       matrix:
@@ -248,7 +277,7 @@ jobs:
           pnpm install
       - name: Install golang-migrate for Clickhouse migrations
         run: |
-          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.16.2/migrate.linux-amd64.tar.gz | tar xvz
+          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.18.2/migrate.linux-amd64.tar.gz | tar xvz
           sudo mv migrate /usr/bin/migrate
           which migrate
       - name: Load default env
@@ -281,6 +310,9 @@ jobs:
 
   e2e-tests:
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     steps:
       - uses: actions/checkout@v4
       - uses: pnpm/action-setup@v3
@@ -322,6 +354,9 @@ jobs:
 
   e2e-server-tests:
     runs-on: ubuntu-latest
+    needs:
+      - pre-job
+    if: needs.pre-job.outputs.should_skip != 'true'
     steps:
       - uses: actions/checkout@v4
       - name: Login to Docker Hub
@@ -343,7 +378,7 @@ jobs:
           pnpm install
       - name: Install golang-migrate for Clickhouse migrations
         run: |
-          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.16.2/migrate.linux-amd64.tar.gz | tar xvz
+          curl -L https://github.com/golang-migrate/migrate/releases/download/v4.18.2/migrate.linux-amd64.tar.gz | tar xvz
           sudo mv migrate /usr/bin/migrate
           which migrate
       - name: Load default env
@@ -459,9 +494,9 @@ jobs:
             type=ref,event=pr
             type=sha
             type=semver,pattern={{version}}
-            type=semver,pattern={{major}}.{{minor}}
-            type=semver,pattern={{major}}
-            type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v3') }}
+            type=semver,pattern={{major}}.{{minor}},enable=${{ !contains(github.ref, '-rc') }}
+            type=semver,pattern={{major}},enable=${{ !contains(github.ref, '-rc') }}
+            type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v3') && !contains(github.ref, '-rc') }}
       - name: Build and push Docker image (web)
         uses: docker/build-push-action@v4
         with:
@@ -487,9 +522,9 @@ jobs:
             type=ref,event=pr
             type=sha
             type=semver,pattern={{version}}
-            type=semver,pattern={{major}}.{{minor}}
-            type=semver,pattern={{major}}
-            type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v3') }}
+            type=semver,pattern={{major}}.{{minor}},enable=${{ !contains(github.ref, '-rc') }}
+            type=semver,pattern={{major}},enable=${{ !contains(github.ref, '-rc') }}
+            type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v3') && !contains(github.ref, '-rc') }}
       - name: Build and push Docker image (worker)
         uses: docker/build-push-action@v4
         with:

.github/workflows/snyk.yml

@@ -6,7 +6,7 @@ on:
   push:
     branches:
       - main
-  merge_group:
+  # Snyk cannot upload results in merge group. Hence, we only run on PRs and when pushingon the main branch https://github.com/github/codeql-action/issues/1572
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
@@ -29,6 +29,13 @@ jobs:
           image: langfuse-langfuse-web
           args: --file=web/Dockerfile
 
+      # Workaround for https://github.com/github/codeql-action/issues/2187
+      - name: Replace security-severity undefined for license-related findings
+        run: "sed -i 's/\"security-severity\": \"undefined\"/\"security-severity\": \"0\"/g' snyk.sarif"
+
+      - name: Replace security-severity null for license-related findings
+        run: "sed -i 's/\"security-severity\": \"null\"/\"security-severity\": \"0\"/g' snyk.sarif"
+
       - name: Upload result to GitHub Code Scanning
         uses: github/codeql-action/upload-sarif@v3
         with:
@@ -44,6 +51,13 @@ jobs:
           image: langfuse-langfuse-worker
           args: --file=worker/Dockerfile
 
+      # Workaround for https://github.com/github/codeql-action/issues/2187
+      - name: Replace security-severity undefined for license-related findings
+        run: "sed -i 's/\"security-severity\": \"undefined\"/\"security-severity\": \"0\"/g' snyk.sarif"
+
+      - name: Replace security-severity null for license-related findings
+        run: "sed -i 's/\"security-severity\": \"null\"/\"security-severity\": \"0\"/g' snyk.sarif"
+
       - name: Upload result to GitHub Code Scanning
         uses: github/codeql-action/upload-sarif@v3
         with:

.gitignore

@@ -39,7 +39,6 @@ yarn-error.log*
 !.env.dev.example
 !.env.dev-azure.example
 !.env.prod.example
-!.env.dev.legacy.example
 
 # vercel
 .vercel

CONTRIBUTING.md

@@ -121,36 +121,40 @@ flowchart TB
         end
     end
 
-    DB[Postgres Database]
-    Redis[Redis Cache/Queue]
-    Clickhouse[Clickhouse Database]
+    subgraph s9 ["VPC (US and EU separated)"]
+        DB[Postgres Database]
+        Redis[Redis Cache/Queue]
+        Clickhouse[Clickhouse Database]
+
+        subgraph s1["Application (langfuse/langfuse/web)"]
+            API[Public HTTP API]
+            G[TRPC API]
+            I[NextAuth]
+            H[React Frontend]
+            ORM
+            H --> G
+            H --> I
+            G --> I
+            G --- ORM
+            API --- ORM
+            I --- ORM
+        end
+
+        subgraph s5["Application (langfuse/langfuse/worker)"]
+            Worker
+        end
+
+        Worker --- DB
+        Worker --- Redis
+        Worker --- Clickhouse
+
+        ORM --- DB
+        ORM --- Redis
+        ORM --- Clickhouse
+
 
-    subgraph s1["Application (langfuse/langfuse/web)"]
-        API[Public HTTP API]
-        G[TRPC API]
-        I[NextAuth]
-        H[React Frontend]
-        ORM
-        H --> G
-        H --> I
-        G --> I
-        G --- ORM
-        API --- ORM
-        I --- ORM
     end
 
-    subgraph s5["Application (langfuse/langfuse/worker)"]
-        Worker
-    end
-
-    Worker --- DB
-    Worker --- Redis
-    Worker --- Clickhouse
-
-    ORM --- DB
-    ORM --- Redis
-    ORM --- Clickhouse
-
     JS --- API
     Python --- API
 ```
@@ -397,7 +401,7 @@ The background color of the following component will be `hsl(var(--primary))` an
 | --primary-accent         | Primary accent color used for branding                             | Layout                           |
 | --hover-primary-accent   | Primary accent color used for hover effects for links              | SignIn and AuthCloudRegionSwitch |
 | --muted-green            | Muted green for Event label                                        | ObservationTree                  |
-| --muted-orange           | Muted orange for Generation label                                  | ObservationTree                  |
+| --muted-magenta          | Muted magenta for Generation label                                 | ObservationTree                  |
 | --muted-blue             | Muted blue for Span label                                          | ObservationTree                  |
 | --muted-gray             | Muted gray for disabled status badges                              | StatusBadge                      |
 | --accent-light-green     | Light green accent for background of output and assistant messages | IOPreview, Generations, Traces   |
@@ -439,7 +443,7 @@ You can update the default AI models and prices by adding or updating an entry i
 Please note that
 
 - prices are in USD
-- the list is ordered by ID, so make sure to keep this order
+- the list is ordered by ID, so make sure to keep this order and insert new models at the end of the list
 - the `updated_at` field must be updated with the current date in ISO 8601 format. Otherwise, the change will be ignored.
 
 ### Transition period until V3 release

README.cn.md

@@ -0,0 +1,354 @@
+![Langfuse GitHub Banner](https://langfuse.com/images/docs/github-readme/github-banner.png)
+
+<div align="center">
+  <div>
+    <h3>
+    <a href="https://langfuse.com/cn">
+        <strong>🇨🇳 🤝 🪢</strong>
+      </a> · 
+      <a href="https://cloud.langfuse.com">
+        <strong>Langfuse Cloud</strong>
+      </a> · 
+      <a href="https://langfuse.com/docs/deployment/self-host">
+        <strong>自托管</strong>
+      </a> · 
+      <a href="https://langfuse.com/demo">
+        <strong>演示</strong>
+      </a>
+    </h3>
+  </div>
+
+  <div>
+    <a href="https://langfuse.com/docs"><strong>文档</strong></a> ·
+    <a href="https://langfuse.com/issues"><strong>报告问题</strong></a> ·
+    <a href="https://langfuse.com/ideas"><strong>功能请求</strong></a> ·
+    <a href="https://langfuse.com/changelog"><strong>更新日志</strong></a> ·
+    <a href="https://langfuse.com/roadmap"><strong>路线图</strong></a> ·
+  </div>
+  <br/>
+  <span>Langfuse 使用 <a href="https://github.com/orgs/langfuse/discussions"><strong>Github Discussions</strong></a> 作为支持和功能请求的平台。</span>
+  <br/>
+  <span><b>我们正在招聘。</b> <a href="https://langfuse.com/careers"><strong>加入我们</strong></a>，从事产品工程和技术市场职位。</span>
+  <br/>
+  <br/>
+  <div>
+  </div>
+</div>
+
+<p align="center">
+  <a href="https://github.com/langfuse/langfuse/blob/main/LICENSE">
+    <img src="https://img.shields.io/badge/License-MIT-E11311.svg" alt="MIT License">
+  </a>
+  <a href="https://www.ycombinator.com/companies/langfuse">
+    <img src="https://img.shields.io/badge/Y%20Combinator-W23-orange" alt="Y Combinator W23">
+  </a>
+  <a href="https://hub.docker.com/u/langfuse" target="_blank">
+    <img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langfuse/langfuse?labelColor=%20%23FDB062&logo=Docker&labelColor=%20%23528bff">
+  </a>
+  <a href="https://pypi.python.org/pypi/langfuse">
+    <img src="https://img.shields.io/pypi/dm/langfuse?logo=python&logoColor=white&label=pypi%20langfuse&color=blue" alt="langfuse Python package on PyPi">
+  </a>
+  <a href="https://www.npmjs.com/package/langfuse">
+    <img src="https://img.shields.io/npm/dm/langfuse?logo=npm&logoColor=white&label=npm%20langfuse&color=blue" alt="langfuse npm package">
+  </a>
+  <br/>
+  <a href="https://discord.com/invite/7NXusRtqYU" target="_blank">
+    <img src="https://img.shields.io/discord/1111061815649124414?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
+         alt="在 Discord 上聊天">
+  </a>
+  <a href="https://twitter.com/intent/follow?screen_name=langfuse" target="_blank">
+    <img src="https://img.shields.io/twitter/follow/langfuse?logo=X&color=%20%23f5f5f5"
+         alt="在 X (Twitter) 上关注">
+  </a>
+  <a href="https://www.linkedin.com/company/langfuse/" target="_blank">
+    <img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
+         alt="在 LinkedIn 上关注">
+  </a>
+  <a href="https://github.com/langfuse/langfuse/graphs/commit-activity" target="_blank">
+    <img alt="过去一个月的提交" src="https://img.shields.io/github/commit-activity/m/langfuse/langfuse?labelColor=%20%2332b583&color=%20%2312b76a">
+  </a>
+  <a href="https://github.com/langfuse/langfuse/" target="_blank">
+    <img alt="已关闭的问题" src="https://img.shields.io/github/issues-search?query=repo%3Alangfuse%2Flangfuse%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98">
+  </a>
+  <a href="https://github.com/langfuse/langfuse/discussions/" target="_blank">
+    <img alt="讨论帖数量" src="https://img.shields.io/github/discussions/langfuse/langfuse?labelColor=%20%239b8afb&color=%20%237a5af8">
+  </a>
+</p>
+
+<p align="center">
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README.cn.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README.ja.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README.kr.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+</p>
+
+Langfuse 是一个 **开源 LLM 工程** 平台。它帮助团队协作 **开发、监控、评估** 以及 **调试** AI 应用。Langfuse 可在几分钟内 **自托管**，并且经过 **实战考验**。
+
+[![Langfuse 概览视频](https://github.com/user-attachments/assets/3926b288-ff61-4b95-8aa1-45d041c70866)](https://langfuse.com/watch-demo)
+
+## ✨ 核心特性
+
+![Langfuse 概览](https://langfuse.com/images/docs/github-readme/github-feature-overview.png)
+
+- [LLM 应用可观察性](https://langfuse.com/docs/tracing)：为你的应用插入仪表代码，并开始将追踪数据传送到 Langfuse，从而追踪 LLM 调用及应用中其他相关逻辑（如检索、嵌入或代理操作）。检查并调试复杂日志及用户会话。试试互动的 [演示](https://langfuse.com/docs/demo) 看看效果。
+  
+- [提示管理](https://langfuse.com/docs/prompts/get-started) 帮助你集中管理、版本控制并协作迭代提示。得益于服务器和客户端的高效缓存，你可以在不增加延迟的情况下反复迭代提示。
+
+- [评估](https://langfuse.com/docs/scores/overview) 是 LLM 应用开发流程的关键组成部分，Langfuse 能够满足你的多样需求。它支持 LLM 作为“裁判”、用户反馈收集、手动标注以及通过 API/SDK 实现自定义评估流程。
+
+- [数据集](https://langfuse.com/docs/datasets/overview) 为评估你的 LLM 应用提供测试集和基准。它们支持持续改进、部署前测试、结构化实验、灵活评估，并能与 LangChain、LlamaIndex 等框架无缝整合。
+
+- [LLM 试玩平台](https://langfuse.com/docs/playground) 是用于测试和迭代提示及模型配置的工具，缩短反馈周期，加速开发。当你在追踪中发现异常结果时，可以直接跳转至试玩平台进行调整。
+
+- [综合 API](https://langfuse.com/docs/api)：Langfuse 常用于驱动定制化的 LLMOps 工作流程，同时利用 Langfuse 提供的构建模块和 API。我们提供 OpenAPI 规格、Postman 集合以及针对 Python 和 JS/TS 的类型化 SDK。
+
+## 📦 部署 Langfuse
+
+![Langfuse 部署选项](https://langfuse.com/images/docs/github-readme/github-deployment-options.png)
+
+### Langfuse Cloud
+
+由 Langfuse 团队管理的部署，提供慷慨的免费额度（爱好者计划），无需信用卡。
+
+<div align="center">
+  <a href="https://cloud.langfuse.com" target="_blank">
+    <img alt="注册 Langfuse Cloud" src="https://img.shields.io/badge/»%20Sign%20up%20for%20Langfuse%20Cloud-8A2BE2?&color=orange">
+  </a>
+</div>
+
+### 自托管 Langfuse
+
+在你自己的基础设施上运行 Langfuse：
+
+- [本地（docker compose）](https://langfuse.com/self-hosting/local)：使用 Docker Compose 在你的机器上于 5 分钟内运行 Langfuse。
+
+  ````bash:README.md/docker-compose
+  # 获取最新的 Langfuse 仓库副本
+  git clone https://github.com/langfuse/langfuse.git
+  cd langfuse
+
+  # 运行 Langfuse 的 docker compose
+  docker compose up
+  `````
+
+- [Kubernetes（Helm）](https://langfuse.com/self-hosting/kubernetes-helm)：使用 Helm 在 Kubernetes 集群上部署 Langfuse。这是推荐的生产环境部署方式。
+
+- [虚拟机](https://langfuse.com/self-hosting/docker-compose)：使用 Docker Compose 在单台虚拟机上部署 Langfuse。
+
+- 【计划中】：针对各云平台的部署指南，欢迎在以下讨论中投票和评论：[AWS](https://github.com/orgs/langfuse/discussions/4645)、[Google Cloud](https://github.com/langfuse/discussions/4646)、[Azure](https://github.com/orgs/langfuse/discussions/4647)。
+
+请参阅 [自托管文档](https://langfuse.com/self-hosting) 了解更多关于架构和配置选项的信息。
+
+## 🔌 集成
+
+![Langfuse 集成](https://langfuse.com/images/docs/github-readme/github-integrations.png)
+
+### 主要集成：
+
+| 集成                                                                   | 支持语言/平台            | 描述                                                                                                                                       |
+| ---------------------------------------------------------------------- | ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------- |
+| [SDK](https://langfuse.com/docs/sdk)                                    | Python, JS/TS            | 使用 SDK 进行手动仪表化，实现全面灵活性。                                                                                                   |
+| [OpenAI](https://langfuse.com/docs/integrations/openai)                 | Python, JS/TS            | 通过直接替换 OpenAI SDK 实现自动仪表化。                                                                                                    |
+| [Langchain](https://langfuse.com/docs/integrations/langchain)           | Python, JS/TS            | 通过传入回调处理器至 Langchain 应用实现自动仪表化。                                                                                         |
+| [LlamaIndex](https://langfuse.com/docs/integrations/llama-index/get-started) | Python                   | 通过 LlamaIndex 回调系统实现自动仪表化。                                                                                                    |
+| [Haystack](https://langfuse.com/docs/integrations/haystack)             | Python                   | 通过 Haystack 内容追踪系统实现自动仪表化。                                                                                                  |
+| [LiteLLM](https://langfuse.com/docs/integrations/litellm)               | Python, JS/TS (仅代理)     | 允许使用任何 LLM 替代 GPT。支持 Azure、OpenAI、Cohere、Anthropic、Ollama、VLLM、Sagemaker、HuggingFace、Replicate（100+ LLMs）。         |
+| [Vercel AI SDK](https://langfuse.com/docs/integrations/vercel-ai-sdk)     | JS/TS                    | 基于 TypeScript 的工具包，帮助开发者使用 React、Next.js、Vue、Svelte 和 Node.js 构建 AI 驱动的应用。                                           |
+| [API](https://langfuse.com/docs/api)                                    |                          | 直接调用公共 API。提供 OpenAPI 规格。                                                                                                       |
+
+### 与 Langfuse 集成的软件包：
+
+| 名称                                                                   | 类型                | 描述                                                                                                     |
+| ---------------------------------------------------------------------- | ------------------- | -------------------------------------------------------------------------------------------------------- |
+| [Instructor](https://langfuse.com/docs/integrations/instructor)         | 库                  | 用于获取结构化 LLM 输出（JSON、Pydantic）的库。                                                         |
+| [DSPy](https://langfuse.com/docs/integrations/dspy)                     | 库                  | 一个系统性优化语言模型提示和权重的框架。                                                                 |
+| [Mirascope](https://langfuse.com/docs/integrations/mirascope)           | 库                  | 构建 LLM 应用的 Python 工具包。                                                                           |
+| [Ollama](https://langfuse.com/docs/integrations/ollama)                 | 模型（本地）         | 在你的机器上轻松运行开源 LLM。                                                                            |
+| [Amazon Bedrock](https://langfuse.com/docs/integrations/amazon-bedrock) | 模型                | 在 AWS 上运行基础模型和微调模型。                                                                         |
+| [AutoGen](https://langfuse.com/docs/integrations/autogen)               | 代理框架             | 用于构建分布式代理的开源 LLM 平台。                                                                       |
+| [Flowise](https://langfuse.com/docs/integrations/flowise)               | 聊天/代理界面         | 基于 JS/TS 的无代码构建器，用于定制化 LLM 流程。                                                          |
+| [Langflow](https://langfuse.com/docs/integrations/langflow)             | 聊天/代理界面         | 基于 Python 的 LangChain 用户界面，采用 react-flow 设计，提供便捷的实验与原型构建体验。                   |
+| [Dify](https://langfuse.com/docs/integrations/dify)                     | 聊天/代理界面         | 带有无代码构建器的开源 LLM 应用开发平台。                                                                 |
+| [OpenWebUI](https://langfuse.com/docs/integrations/openwebui)           | 聊天/代理界面         | 自托管的 LLM 聊天网页界面，支持包括自托管和本地模型在内的多种 LLM 运行器。                                |
+| [Promptfoo](https://langfuse.com/docs/integrations/promptfoo)           | 工具                | 开源 LLM 测试平台。                                                                                       |
+| [LobeChat](https://langfuse.com/docs/integrations/lobechat)             | 聊天/代理界面         | 开源聊天机器人平台。                                                                                       |
+| [Vapi](https://langfuse.com/docs/integrations/vapi)                     | 平台                | 开源语音 AI 平台。                                                                                        |
+| [Inferable](https://langfuse.com/docs/integrations/other/inferable)     | 代理                | 构建分布式代理的开源 LLM 平台。                                                                            |
+| [Gradio](https://langfuse.com/docs/integrations/other/gradio)           | 聊天/代理界面         | 开源 Python 库，可用于构建类似聊天 UI 的网页界面。                                                        |
+| [Goose](https://langfuse.com/docs/integrations/goose)                   | 代理                | 构建分布式代理的开源 LLM 平台。                                                                            |
+| [smolagents](https://langfuse.com/docs/integrations/smolagents)         | 代理                | 开源 AI 代理框架。                                                                                        |
+| [CrewAI](https://langfuse.com/docs/integrations/crewai)                 | 代理                | 多代理框架，用于实现代理之间的协作与工具调用。                                                           |
+
+## 🚀 快速入门
+
+为你的应用增加仪表代码，并开始将追踪数据上传到 Langfuse，从而记录 LLM 调用及应用中其他相关逻辑（如检索、嵌入或代理操作）。
+
+### 1️⃣ 创建新项目
+
+1. [创建 Langfuse 账户](https://cloud.langfuse.com/auth/sign-up) 或 [自托管](https://langfuse.com/self-hosting)
+2. 创建一个新项目
+3. 在项目设置中创建新的 API 凭证
+
+### 2️⃣ 记录你的第一个 LLM 调用
+
+使用 [<code>@observe()</code> 装饰器](https://langfuse.com/docs/sdk/python/decorators) 可轻松跟踪任何 Python LLM 应用。在本快速入门中，我们还使用了 Langfuse 的 [OpenAI 集成](https://langfuse.com/docs/integrations/openai) 来自动捕获所有模型参数。
+
+> [!提示]
+> 不使用 OpenAI？请访问 [我们的文档](https://langfuse.com/docs/get-started#log-your-first-llm-call-to-langfuse) 了解如何记录其他模型和框架。
+
+安装依赖：
+
+````bash
+pip install langfuse openai
+````
+
+配置环境变量（创建名为 **.env** 的文件）：
+
+````bash:.env
+LANGFUSE_SECRET_KEY="sk-lf-..."
+LANGFUSE_PUBLIC_KEY="pk-lf-..."
+LANGFUSE_HOST="https://cloud.langfuse.com" # 🇪🇺 欧盟区域
+# LANGFUSE_HOST="https://us.cloud.langfuse.com" # 🇺🇸 美洲区域
+````
+
+创建示例代码（文件名：**main.py**）：
+
+````python:main.py
+from langfuse.decorators import observe
+from langfuse.openai import openai  # OpenAI 集成
+
+@observe()
+def story():
+    return openai.chat.completions.create(
+        model="gpt-4o",
+        messages=[{"role": "user", "content": "What is Langfuse?"}],
+    ).choices[0].message.content
+
+@observe()
+def main():
+    return story()
+
+main()
+````
+
+### 3️⃣ 在 Langfuse 中查看追踪记录
+
+在 Langfuse 中查看你的语言模型调用及其他应用逻辑。
+
+![示例追踪记录](https://langfuse.com/images/docs/github-readme/github-example-trace.png)
+
+_[Langfuse 中的公共示例追踪](https://cloud.langfuse.com/project/cloramnkj0002jz088vzn1ja4/traces/2cec01e3-3dc2-472f-afcf-3b968cf0c1f4?timestamp=2025-02-10T14%3A27%3A30.275Z&observation=cb5ff844-07ef-41e6-b8e2-6c64344bc13b)_
+
+> [!提示]
+>
+> [了解更多](https://langfuse.com/docs/tracing) 关于 Langfuse 中的追踪，或试试 [互动演示](https://langfuse.com/docs/demo)。
+
+## ⭐️ 给我们加星
+
+![为 Langfuse 加星](https://langfuse.com/images/docs/github-readme/github-star-howto.gif)
+
+## 💭 支持
+
+查找问题答案：
+
+- 我们的 [文档](https://langfuse.com/docs) 是查找答案的最佳起点。内容全面，我们投入大量时间进行维护。你也可以通过 GitHub 提出文档修改建议。
+- [Langfuse 常见问题](https://langfuse.com/faq) 解答了最常见的问题。
+- 使用 “[Ask AI](https://langfuse.com/docs/ask-ai)” 立即获取问题答案。
+
+支持渠道：
+
+- **在 GitHub Discussions 的 [公共问答](https://github.com/orgs/langfuse/discussions/categories/support) 中提出任何问题。** 请尽量提供详细信息（如代码片段、截图、背景信息）以帮助我们理解你的问题。
+- 在 GitHub Discussions 中 [提出功能请求](https://github.com/orgs/langfuse/discussions/categories/ideas)。
+- 在 GitHub Issues 中 [报告 Bug](https://github.com/langfuse/langfuse/issues)。
+- 对于时效性较强的问题，请通过应用内聊天小部件联系我们。
+
+## 🤝 贡献
+
+欢迎你的贡献！
+
+- 在 GitHub Discussions 中为 [想法](https://github.com/orgs/langfuse/discussions/categories/ideas)投票。
+- 提出并评论 [问题](https://github.com/langfuse/langfuse/issues)。
+- 提交 PR —— 详情请参见 [CONTRIBUTING.md](CONTRIBUTING.md)，了解如何搭建开发环境。
+
+## 🥇 许可证
+
+除 `ee` 文件夹外，本仓库采用 MIT 许可证。详情请参见 [LICENSE](LICENSE) 以及 [文档](https://langfuse.com/docs/open-source)。
+
+## ⭐️ 星标历史
+
+<a href="https://star-history.com/#langfuse/langfuse&Date">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date&theme=dark" />
+    <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
+    <img alt="星标历史图表" src="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" style="border-radius: 15px;" />
+  </picture>
+</a>
+
+## ❤️ 使用 Langfuse 的开源项目
+
+以下是使用 Langfuse 的顶级开源 Python 项目，按星标数排名（[来源](https://github.com/langfuse/langfuse-docs/blob/main/components-mdx/dependents)）：
+
+| 仓库                                                                                                                                                                                                                                                             | 星数  |
+| :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----: |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/127165244?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langgenius](https://github.com/langgenius) / [dify](https://github.com/langgenius/dify)                                      | 54865 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [open-webui](https://github.com/open-webui/open-webui)                            | 51531 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/131470832?s=40&v=4" width="20" height="20" alt=""> &nbsp; [lobehub](https://github.com/lobehub) / [lobe-chat](https://github.com/lobehub/lobe-chat)                                   | 49003 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85702467?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langflow-ai](https://github.com/langflow-ai) / [langflow](https://github.com/langflow-ai/langflow)                              | 39093 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/130722866?s=40&v=4" width="20" height="20" alt=""> &nbsp; [run-llama](https://github.com/run-llama) / [llama_index](https://github.com/run-llama/llama_index)                                 | 37368 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/139558948?s=40&v=4" width="20" height="20" alt=""> &nbsp; [chatchat-space](https://github.com/chatchat-space) / [Langchain-Chatchat](https://github.com/chatchat-space/Langchain-Chatchat)    | 32486 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/128289781?s=40&v=4" width="20" height="20" alt=""> &nbsp; [FlowiseAI](https://github.com/FlowiseAI) / [Flowise](https://github.com/FlowiseAI/Flowise)                                      | 32448 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/31035808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mindsdb](https://github.com/mindsdb) / [mindsdb](https://github.com/mindsdb/mindsdb)                                             | 26931 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/119600397?s=40&v=4" width="20" height="20" alt=""> &nbsp; [twentyhq](https://github.com/twentyhq) / [twenty](https://github.com/twentyhq/twenty)                                   | 24195 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog](https://github.com/PostHog/posthog)                                             | 22618 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/121462774?s=40&v=4" width="20" height="20" alt=""> &nbsp; [BerriAI](https://github.com/BerriAI) / [litellm](https://github.com/BerriAI/litellm)                                    | 15151 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/179202840?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mediar-ai](https://github.com/mediar-ai) / [screenpipe](https://github.com/mediar-ai/screenpipe)                            | 11037 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/105877416?s=40&v=4" width="20" height="20" alt=""> &nbsp; [formbricks](https://github.com/formbricks) / [formbricks](https://github.com/formbricks/formbricks)                         |  9386 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/76263028?s=40&v=4" width="20" height="20" alt=""> &nbsp; [anthropics](https://github.com/anthropics) / [courses](https://github.com/anthropics/courses)                                |  8385 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/78410652?s=40&v=4" width="20" height="20" alt=""> &nbsp; [GreyDGL](https://github.com/GreyDGL) / [PentestGPT](https://github.com/GreyDGL/PentestGPT)                           |  7374 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/152537519?s=40&v=4" width="20" height="20" alt=""> &nbsp; [superagent-ai](https://github.com/superagent-ai) / [superagent](https://github.com/superagent-ai/superagent)               |  5391 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/137907881?s=40&v=4" width="20" height="20" alt=""> &nbsp; [promptfoo](https://github.com/promptfoo) / [promptfoo](https://github.com/promptfoo/promptfoo)                           |  4976 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/157326433?s=40&v=4" width="20" height="20" alt=""> &nbsp; [onlook-dev](https://github.com/onlook-dev) / [onlook](https://github.com/onlook-dev/onlook)                                  |  4141 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/7250217?s=40&v=4" width="20" height="20" alt=""> &nbsp; [Canner](https://github.com/Canner) / [WrenAI](https://github.com/Canner/WrenAI)                                      |  2526 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/11855343?s=40&v=4" width="20" height="20" alt=""> &nbsp; [pingcap](https://github.com/pingcap) / [autoflow](https://github.com/pingcap/autoflow)                                  |  2061 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85268109?s=40&v=4" width="20" height="20" alt=""> &nbsp; [MLSysOps](https://github.com/MLSysOps) / [MLE-agent](https://github.com/MLSysOps/MLE-agent)                         |  1161 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [pipelines](https://github.com/open-webui/pipelines)                          |  1100 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/18422723?s=40&v=4" width="20" height="20" alt=""> &nbsp; [alishobeiri](https://github.com/alishobeiri) / [thread](https://github.com/alishobeiri/thread)                           |  1074 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/125468716?s=40&v=4" width="20" height="20" alt=""> &nbsp; [topoteretes](https://github.com/topoteretes) / [cognee](https://github.com/topoteretes/cognee)                          |   971 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/188657705?s=40&v=4" width="20" height="20" alt=""> &nbsp; [bRAGAI](https://github.com/bRAGAI) / [bRAG-langchain](https://github.com/bRAGAI/bRAG-langchain)                         |   823 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169500408?s=40&v=4" width="20" height="20" alt=""> &nbsp; [opslane](https://github.com/opslane) / [opslane](https://github.com/opslane/opslane)                             |   677 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/151867818?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dynamiq-ai](https://github.com/dynamiq-ai) / [dynamiq](https://github.com/dynamiq-ai/dynamiq)                         |   639 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/48585267?s=40&v=4" width="20" height="20" alt=""> &nbsp; [theopenconversationkit](https://github.com/theopenconversationkit) / [tock](https://github.com/theopenconversationkit/tock) |   514 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/20493493?s=40&v=4" width="20" height="20" alt=""> &nbsp; [andysingal](https://github.com/andysingal) / [llm-course](https://github.com/andysingal/llm-course)                         |   394 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/132396805?s=40&v=4" width="20" height="20" alt=""> &nbsp; [phospho-app](https://github.com/phospho-app) / [phospho](https://github.com/phospho-app/phospho)                           |   384 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/178644984?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sentient-engineering](https://github.com/sentient-engineering) / [agent-q](https://github.com/sentient-engineering/agent-q) |   370 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/168552753?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sql-agi](https://github.com/sql-agi) / [DB-GPT](https://github.com/sql-agi/DB-GPT)                          |   324 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog-foss](https://github.com/PostHog/posthog-foss)                           |   305 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/154247157?s=40&v=4" width="20" height="20" alt=""> &nbsp; [vespperhq](https://github.com/vespperhq) / [vespper](https://github.com/vespperhq/vespper)                         |   304 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/185116535?s=40&v=4" width="20" height="20" alt=""> &nbsp; [block](https://github.com/block) / [goose](https://github.com/block/goose)                                  |   295 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/609489?s=40&v=4" width="20" height="20" alt=""> &nbsp; [aorwall](https://github.com/aorwall) / [moatless-tools](https://github.com/aorwall/moatless-tools)                 |   291 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/2357342?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dmayboroda](https://github.com/dmayboroda) / [minima](https://github.com/dmayboroda/minima)                          |   221 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/66303003?s=40&v=4" width="20" height="20" alt=""> &nbsp; [RobotecAI](https://github.com/RobotecAI) / [rai](https://github.com/RobotecAI/rai)                              |   172 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/148684274?s=40&v=4" width="20" height="20" alt=""> &nbsp; [i-am-alice](https://github.com/i-am-alice) / [3rd-devs](https://github.com/i-am-alice/3rd-devs)                   |   148 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/171735272?s=40&v=4" width="20" height="20" alt=""> &nbsp; [8090-inc](https://github.com/8090-inc) / [xrx-sample-apps](https://github.com/8090-inc/xrx-sample-apps)       |   138 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/104478511?s=40&v=4" width="20" height="20" alt=""> &nbsp; [babelcloud](https://github.com/babelcloud) / [LLM-RGB](https://github.com/babelcloud/LLM-RGB)                  |   135 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/15125613?s=40&v=4" width="20" height="20" alt=""> &nbsp; [souzatharsis](https://github.com/souzatharsis) / [tamingLLMs](https://github.com/souzatharsis/tamingLLMs)        |   129 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169401942?s=40&v=4" width="20" height="20" alt=""> &nbsp; [LibreChat-AI](https://github.com/LibreChat-AI) / [librechat.ai](https://github.com/LibreChat-AI/librechat.ai)   |   128 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/51827949?s=40&v=4" width="20" height="20" alt=""> &nbsp; [deepset-ai](https://github.com/deepset-ai) / [haystack-core-integrations](https://github.com/deepset-ai/haystack-core-integrations) |   126 |
+
+## 🔒 安全与隐私
+
+我们非常重视数据安全和隐私。更多信息请参阅我们的 [安全与隐私](https://langfuse.com/security) 页面。
+
+### 遥测
+
+默认情况下，Langfuse 会自动将自托管实例的基础使用统计数据上传至集中服务器（PostHog）。
+
+这有助于我们：
+
+1. 了解 Langfuse 的使用情况，并改进最关键的功能。
+2. 跟踪整体使用数据，以便内部及外部（例如筹款）报告。
+
+所有数据均不会与第三方共享，也不包含任何敏感信息。我们对这一过程保持高度透明，你可以在 [此处](/web/src/features/telemetry/index.ts) 查看我们收集的具体数据。
+
+你可以通过设置 `TELEMETRY_ENABLED=false` 来选择退出。
+````

README.ja.md

@@ -0,0 +1,361 @@
+![Langfuse GitHub Banner](https://langfuse.com/images/docs/github-readme/github-banner.png)
+
+<div align="center">
+   <div>
+      <h3>
+          <a href="https://langfuse.com/jp">
+            <strong>🇯🇵 🤝 🪢</strong>
+         </a> · 
+         <a href="https://cloud.langfuse.com">
+            <strong>Langfuse Cloud</strong>
+         </a> · 
+         <a href="https://langfuse.com/docs/deployment/self-host">
+            <strong>セルフホスティング</strong>
+         </a> · 
+         <a href="https://langfuse.com/demo">
+            <strong>デモ</strong>
+         </a>
+      </h3>
+   </div>
+
+   <div>
+      <a href="https://langfuse.com/docs"><strong>ドキュメント</strong></a> ·
+      <a href="https://langfuse.com/issues"><strong>バグ報告</strong></a> ·
+      <a href="https://langfuse.com/ideas"><strong>機能リクエスト</strong></a> ·
+      <a href="https://langfuse.com/changelog"><strong>変更履歴</strong></a> ·
+      <a href="https://langfuse.com/roadmap"><strong>ロードマップ</strong></a> ·
+   </div>
+   <br/>
+   <span>Langfuseは、サポートと機能リクエストのために <a href="https://github.com/orgs/langfuse/discussions"><strong>Github Discussions</strong></a> を利用しています。</span>
+   <br/>
+   <span><b>採用中です。</b> <a href="https://langfuse.com/careers"><strong>当社で働く</strong></a> （製品エンジニアリングおよびテクニカル・ゴートゥーマーケットのポジション）への応募をお待ちしています。</span>
+   <br/>
+   <br/>
+   <div>
+   </div>
+</div>
+
+<p align="center">
+   <a href="https://github.com/langfuse/langfuse/blob/main/LICENSE">
+   <img src="https://img.shields.io/badge/License-MIT-E11311.svg" alt="MIT License">
+   </a>
+   <a href="https://www.ycombinator.com/companies/langfuse"><img src="https://img.shields.io/badge/Y%20Combinator-W23-orange" alt="Y Combinator W23"></a>
+   <a href="https://hub.docker.com/u/langfuse" target="_blank">
+   <img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langfuse/langfuse?labelColor=%20%23FDB062&logo=Docker&labelColor=%20%23528bff"></a>
+   <a href="https://pypi.python.org/pypi/langfuse"><img src="https://img.shields.io/pypi/dm/langfuse?logo=python&logoColor=white&label=pypi%20langfuse&color=blue" alt="langfuse Python package on PyPi"></a>
+   <a href="https://www.npmjs.com/package/langfuse"><img src="https://img.shields.io/npm/dm/langfuse?logo=npm&logoColor=white&label=npm%20langfuse&color=blue" alt="langfuse npm package"></a>
+   <br/>
+   <a href="https://discord.com/invite/7NXusRtqYU" target="_blank">
+   <img src="https://img.shields.io/discord/1111061815649124414?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
+      alt="chat on Discord"></a>
+   <a href="https://twitter.com/intent/follow?screen_name=langfuse" target="_blank">
+   <img src="https://img.shields.io/twitter/follow/langfuse?logo=X&color=%20%23f5f5f5"
+      alt="follow on X(Twitter)"></a>
+   <a href="https://www.linkedin.com/company/langfuse/" target="_blank">
+   <img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
+      alt="follow on LinkedIn"></a>
+   <a href="https://github.com/langfuse/langfuse/graphs/commit-activity" target="_blank">
+   <img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langfuse/langfuse?labelColor=%20%2332b583&color=%20%2312b76a"></a>
+   <a href="https://github.com/langfuse/langfuse/" target="_blank">
+   <img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alangfuse%2Flangfuse%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
+   <a href="https://github.com/langfuse/langfuse/discussions/" target="_blank">
+   <img alt="Discussion posts" src="https://img.shields.io/github/discussions/langfuse/langfuse?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
+</p>
+
+<p align="center">
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README.cn.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README.ja.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README.kr.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+</p>
+
+Langfuseは**オープンソースのLLMエンジニアリング**プラットフォームです。  
+チームが共同でAIアプリケーションを**開発、監視、評価**、および**デバッグ**するのを支援します。  
+Langfuseは**数分でセルフホスト可能**で、**実績に裏打ちされた**システムです。
+
+[![Langfuse Overview Video](https://github.com/user-attachments/assets/3926b288-ff61-4b95-8aa1-45d041c70866)](https://langfuse.com/watch-demo)
+
+## ✨ コア機能
+
+![Langfuse Overview](https://langfuse.com/images/docs/github-readme/github-feature-overview.png)
+
+- **[LLMアプリケーションの可観測性](https://langfuse.com/docs/tracing):**  
+  アプリケーションにインストゥルメンテーションを導入し、Langfuseへトレースを取り込むことで、LLM呼び出しやリトリーバル、埋め込み、エージェントアクションなどの関連ロジックを追跡できます。  
+  複雑なログやユーザーセッションを解析・デバッグできます。  
+  インタラクティブな[デモ](https://langfuse.com/docs/demo)で動作を確認してください。
+
+- **[プロンプト管理](https://langfuse.com/docs/prompts/get-started):**  
+  プロンプトを一元管理し、バージョン管理しながら共同で改善を行えます。  
+  サーバーおよびクライアント側で強力なキャッシングを行うため、アプリケーションのレイテンシを増やすことなくプロンプトの改良が可能です。
+
+- **[評価](https://langfuse.com/docs/scores/overview):**  
+  評価はLLMアプリケーション開発ワークフローの要であり、Langfuseは多様なニーズに対応します。  
+  LLMを判定者として用いる方法、ユーザーフィードバックの収集、手動によるラベリング、API/SDKを通じたカスタム評価パイプラインをサポートします。
+
+- **[データセット](https://langfuse.com/docs/datasets/overview):**  
+  LLMアプリケーション評価用のテストセットやベンチマークを構築できます。  
+  継続的な改善、事前デプロイテスト、構造化された実験、柔軟な評価、さらにLangChainやLlamaIndexなどとのシームレスな統合をサポートします。
+
+- **[LLMプレイグラウンド](https://langfuse.com/docs/playground):**  
+  プロンプトやモデル設定のテスト・反復作業を支援するツールで、フィードバックループを短縮し開発を加速します。  
+  トレースで不具合が見つかった場合、直接プレイグラウンドへ飛び、迅速に改善できます。
+
+- **[包括的なAPI](https://langfuse.com/docs/api):**  
+  LangfuseはAPIを通じて提供されるビルディングブロックを用い、カスタムLLMOpsワークフローの基盤として頻繁に利用されます。  
+  OpenAPI仕様、Postmanコレクション、PythonやJS/TS向けの型付きSDKが利用可能です。
+
+## 📦 Langfuseのデプロイ
+
+![Langfuse Deployment Options](https://langfuse.com/images/docs/github-readme/github-deployment-options.png)
+
+### Langfuse Cloud
+
+Langfuseチームによるマネージドデプロイメント。充実した無料プラン（ホビープラン）で、クレジットカード不要です。
+
+<div align="center">
+    <a href="https://cloud.langfuse.com" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/»%20Sign%20up%20for%20Langfuse%20Cloud-8A2BE2?&color=orange">
+    </a>
+</div>
+
+### セルフホスティング Langfuse
+
+自身のインフラ上でLangfuseを実行できます:
+
+- **[Local (docker compose)](https://langfuse.com/self-hosting/local):**  
+  Docker Composeを使用して、たった5分で自分のマシン上でLangfuseを実行できます.
+
+  ```bash
+  # 最新のLangfuseリポジトリのコピーを取得
+  git clone https://github.com/langfuse/langfuse.git
+  cd langfuse
+
+  # Langfuseのdocker composeを起動
+  docker compose up
+  ```
+
+- **[Kubernetes (Helm)](https://langfuse.com/self-hosting/kubernetes-helm):**  
+  Helmを使用してKubernetesクラスター上でLangfuseを実行します。  
+  こちらが推奨される本番環境でのデプロイ方法です。
+
+- **[VM](https://langfuse.com/self-hosting/docker-compose):**  
+  Docker Composeを使用して、単一の仮想マシン上でLangfuseを実行します。
+
+- **Planned:**  
+  クラウド固有のデプロイガイドは計画中です。以下のスレッドに対して投票やコメントをお願いします:  
+  [AWS](https://github.com/orgs/langfuse/discussions/4645), [Google Cloud](https://github.com/orgs/langfuse/discussions/4646), [Azure](https://github.com/orgs/langfuse/discussions/4647).
+
+[セルフホスティングのドキュメント](https://langfuse.com/self-hosting)を参照し、アーキテクチャや設定オプションの詳細をご確認ください。
+
+## 🔌 インテグレーション
+
+![Langfuse Integrations](https://langfuse.com/images/docs/github-readme/github-integrations.png)
+
+### 主なインテグレーション:
+
+| インテグレーション                                                                  | 対応言語・環境            | 説明                                                                                                                                           |
+| ----------------------------------------------------------------------------------- | ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
+| [SDK](https://langfuse.com/docs/sdk)                                                  | Python, JS/TS             | SDKを利用して手動でインストゥルメンテーションを実装し、完全な柔軟性を提供します。                                                                  |
+| [OpenAI](https://langfuse.com/docs/integrations/openai)                               | Python, JS/TS             | OpenAI SDKのドロップイン置換による自動インストゥルメンテーションを実現します。                                                                   |
+| [Langchain](https://langfuse.com/docs/integrations/langchain)                         | Python, JS/TS             | Langchainアプリケーションにコールバックハンドラーを渡すことで自動的に計測します。                                                                 |
+| [LlamaIndex](https://langfuse.com/docs/integrations/llama-index/get-started)           | Python                    | LlamaIndexのコールバックシステムを介して自動的にインストゥルメントします。                                                                         |
+| [Haystack](https://langfuse.com/docs/integrations/haystack)                           | Python                    | Haystackのコンテンツトレースシステムを利用した自動インストゥルメンテーションを実現します。                                                          |
+| [LiteLLM](https://langfuse.com/docs/integrations/litellm)                             | Python, JS/TS (proxy only)| GPTのドロップイン置換として任意のLLMを使用できます。Azure、OpenAI、Cohere、Anthropic、Ollama、VLLM、Sagemaker、HuggingFace、Replicate（100+ LLM）に対応。 |
+| [Vercel AI SDK](https://langfuse.com/docs/integrations/vercel-ai-sdk)                   | JS/TS                     | React、Next.js、Vue、Svelte、Node.jsを使用してAI搭載アプリケーションの構築を支援するTypeScriptツールキットです。                                      |
+| [API](https://langfuse.com/docs/api)                                                  |                           | 公開APIを直接呼び出すことが可能です。OpenAPI仕様も利用できます。                                                                                 |
+
+### Langfuseと統合されているパッケージ:
+
+| 名前                                                                      | タイプ             | 説明                                                                                                      |
+| ------------------------------------------------------------------------- | ------------------ | --------------------------------------------------------------------------------------------------------- |
+| [Instructor](https://langfuse.com/docs/integrations/instructor)           | ライブラリ         | 構造化されたLLM出力（JSON、Pydantic）を取得するためのライブラリ                                            |
+| [DSPy](https://langfuse.com/docs/integrations/dspy)                       | ライブラリ         | LLMプロンプトや重み付けを体系的に最適化するためのフレームワーク                                             |
+| [Mirascope](https://langfuse.com/docs/integrations/mirascope)             | ライブラリ         | LLMアプリケーション構築用のPythonツールキット                                                               |
+| [Ollama](https://langfuse.com/docs/integrations/ollama)                   | モデル（ローカル） | オープンソースLLMを手軽にローカルで実行するためのツール                                                     |
+| [Amazon Bedrock](https://langfuse.com/docs/integrations/amazon-bedrock)     | モデル             | AWS上でファウンデーションモデルやファインチューニング済みモデルを実行                                         |
+| [AutoGen](https://langfuse.com/docs/integrations/autogen)                 | エージェントフレームワーク | 分散型エージェント構築のためのオープンソースLLMプラットフォーム                                               |
+| [Flowise](https://langfuse.com/docs/integrations/flowise)                 | チャット/エージェント UI | JS/TSのノーコードビルダーで、カスタマイズ可能なLLMフローを構築                                              |
+| [Langflow](https://langfuse.com/docs/integrations/langflow)               | チャット/エージェント UI | PythonベースのUIで、react-flowを用いてLangChainの実験やプロトタイピングを容易に実現                           |
+| [Dify](https://langfuse.com/docs/integrations/dify)                       | チャット/エージェント UI | ノーコードでLLMアプリ開発が可能なオープンソースプラットフォーム                                               |
+| [OpenWebUI](https://langfuse.com/docs/integrations/openwebui)             | チャット/エージェント UI | 自前ホストおよびローカルモデルに対応するLLMチャットWeb UI                                                    |
+| [Promptfoo](https://langfuse.com/docs/integrations/promptfoo)             | ツール             | オープンソースのLLMテストプラットフォーム                                                                   |
+| [LobeChat](https://langfuse.com/docs/integrations/lobechat)               | チャット/エージェント UI | オープンソースのチャットボットプラットフォーム                                                               |
+| [Vapi](https://langfuse.com/docs/integrations/vapi)                       | プラットフォーム   | オープンソースの音声AIプラットフォーム                                                                       |
+| [Inferable](https://langfuse.com/docs/integrations/other/inferable)        | エージェント       | 分散型エージェント構築のためのオープンソースLLMプラットフォーム                                               |
+| [Gradio](https://langfuse.com/docs/integrations/other/gradio)             | チャット/エージェント UI | チャットUIなどのWebインターフェース構築のためのオープンソースPythonライブラリ                                |
+| [Goose](https://langfuse.com/docs/integrations/goose)                   | エージェント       | 分散型エージェント構築のためのオープンソースLLMプラットフォーム                                               |
+| [smolagents](https://langfuse.com/docs/integrations/smolagents)           | エージェント       | オープンソースのAIエージェントフレームワーク                                                                 |
+| [CrewAI](https://langfuse.com/docs/integrations/crewai)                   | エージェント       | エージェントの協調とツール利用を実現するマルチエージェントフレームワーク                                      |
+
+## 🚀 クイックスタート
+
+アプリケーションにインストゥルメンテーションを導入し、LLM呼び出しやリトリーバル、埋め込み、エージェントアクションなどの動作をLangfuseに記録しましょう。  
+複雑なログやユーザーセッションの解析・デバッグが可能になります。
+
+### 1️⃣ 新規プロジェクトの作成
+
+1. [Langfuseアカウント作成](https://cloud.langfuse.com/auth/sign-up) または [セルフホスト](https://langfuse.com/self-hosting)
+2. 新規プロジェクトを作成
+3. プロジェクト設定で新しいAPIクレデンシャルを作成
+
+### 2️⃣ 初めてのLLM呼び出しのログ記録
+
+[`@observe()` デコレーター](https://langfuse.com/docs/sdk/python/decorators)を利用することで、任意のPython製LLMアプリケーションのトレースが簡単に行えます。  
+このクイックスタートでは、Langfuseの[OpenAI統合](https://langfuse.com/docs/integrations/openai)を使用して、全てのモデルパラメータを自動で取得します。
+
+> [!TIP]
+> OpenAIを利用していない場合は、[こちらのドキュメント](https://langfuse.com/docs/get-started#log-your-first-llm-call-to-langfuse)で、他のモデルやフレームワークのログ記録方法をご確認ください。
+
+```bash
+pip install langfuse openai
+```
+
+```bash filename=".env"
+LANGFUSE_SECRET_KEY="sk-lf-..."
+LANGFUSE_PUBLIC_KEY="pk-lf-..."
+LANGFUSE_HOST="https://cloud.langfuse.com" # 🇪🇺 EUリージョン
+# LANGFUSE_HOST="https://us.cloud.langfuse.com" # 🇺🇸 USリージョン
+```
+
+```python:/@observe()/ /from langfuse.openai import openai/ filename="main.py"
+from langfuse.decorators import observe
+from langfuse.openai import openai  # OpenAI統合
+
+@observe()
+def story():
+    return openai.chat.completions.create(
+        model="gpt-4o",
+        messages=[{"role": "user", "content": "What is Langfuse?"}],
+    ).choices[0].message.content
+
+@observe()
+def main():
+    return story()
+
+main()
+```
+
+### 3️⃣ Langfuseでトレースを確認する
+
+Langfuse上で、LLM呼び出しおよびその他のアプリケーションロジックのトレースを確認できます。
+
+![Example trace in Langfuse](https://langfuse.com/images/docs/github-readme/github-example-trace.png)
+
+_[Langfuseの公開トレース例](https://cloud.langfuse.com/project/cloramnkj0002jz088vzn1ja4/traces/2cec01e3-3dc2-472f-afcf-3b968cf0c1f4?timestamp=2025-02-10T14%3A27%3A30.275Z&observation=cb5ff844-07ef-41e6-b8e2-6c64344bc13b)_
+
+> [!TIP]
+>
+> Langfuseでのトレースの詳細については、[こちら](https://langfuse.com/docs/tracing)をご参照いただくか、[インタラクティブデモ](https://langfuse.com/docs/demo)でお試しください。
+
+## ⭐️ Star Langfuse
+
+![Star Langfuse](https://langfuse.com/images/docs/github-readme/github-star-howto.gif)
+
+## 💭 サポート
+
+質問の回答をお探しの場合は:
+
+- 当社の[ドキュメント](https://langfuse.com/docs)は、回答を探すための最良の出発点です。内容が充実しており、継続的なメンテナンスに努めています。GitHubを通じてドキュメントへの修正提案も可能です。
+- よくある質問は[Langfuse FAQ](https://langfuse.com/faq)にまとめられています。
+- [Ask AI](https://langfuse.com/docs/ask-ai)を利用すれば、質問に対して即座に回答を得ることができます。
+
+サポートチャネル:
+
+- **GitHub Discussionsの[パブリックQ&A](https://github.com/orgs/langfuse/discussions/categories/support)で質問してください。**  
+  質問には、コードスニペット、スクリーンショット、背景情報など、できるだけ詳細な情報を含めるとスムーズな対応が可能です。
+- GitHub Discussionsで[機能リクエスト](https://github.com/orgs/langfuse/discussions/categories/ideas)を投稿してください。
+- GitHub Issuesにて[バグ報告](https://github.com/langfuse/langfuse/issues)を行ってください。
+- 緊急の問い合わせの場合は、アプリ内チャットウィジェットでご連絡ください。
+
+## 🤝 貢献
+
+皆様からの貢献を歓迎します!
+
+- GitHub Discussionsの[アイデア](https://github.com/orgs/langfuse/discussions/categories/ideas)に投票してください。
+- [Issues](https://github.com/langfuse/langfuse/issues)を作成・コメントしてください。
+- プルリクエストを送信してください。開発環境のセットアップ方法については[CONTRIBUTING.md](CONTRIBUTING.md)をご参照ください。
+
+## 🥇 ライセンス
+
+このリポジトリは、`ee`フォルダを除き、MITライセンスの下で公開されています。  
+詳細は[LICENSE](LICENSE)および[オープンソースに関するドキュメント](https://langfuse.com/docs/open-source)をご確認ください。
+
+## ⭐️ スターの履歴
+
+<a href="https://star-history.com/#langfuse/langfuse&Date">
+ <picture>
+   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date&theme=dark" />
+   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
+   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" style="border-radius: 15px;" />
+ </picture>
+</a>
+
+## ❤️ Langfuseを利用しているオープンソースプロジェクト
+
+Langfuseを利用している主要なオープンソースPythonプロジェクト（スター数順）: ([出典](https://github.com/langfuse/langfuse-docs/blob/main/components-mdx/dependents))
+
+| リポジトリ                                                                                                                                                                                                                                                          | スター |
+| :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----: |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/127165244?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langgenius](https://github.com/langgenius) / [dify](https://github.com/langgenius/dify)                                            | 54865 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [open-webui](https://github.com/open-webui/open-webui)                                | 51531 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/131470832?s=40&v=4" width="20" height="20" alt=""> &nbsp; [lobehub](https://github.com/lobehub) / [lobe-chat](https://github.com/lobehub/lobe-chat)                                           | 49003 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85702467?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langflow-ai](https://github.com/langflow-ai) / [langflow](https://github.com/langflow-ai/langflow)                                  | 39093 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/130722866?s=40&v=4" width="20" height="20" alt=""> &nbsp; [run-llama](https://github.com/run-llama) / [llama_index](https://github.com/run-llama/llama_index)                                 | 37368 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/139558948?s=40&v=4" width="20" height="20" alt=""> &nbsp; [chatchat-space](https://github.com/chatchat-space) / [Langchain-Chatchat](https://github.com/chatchat-space/Langchain-Chatchat)    | 32486 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/128289781?s=40&v=4" width="20" height="20" alt=""> &nbsp; [FlowiseAI](https://github.com/FlowiseAI) / [Flowise](https://github.com/FlowiseAI/Flowise)                                         | 32448 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/31035808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mindsdb](https://github.com/mindsdb) / [mindsdb](https://github.com/mindsdb/mindsdb)                                                | 26931 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/119600397?s=40&v=4" width="20" height="20" alt=""> &nbsp; [twentyhq](https://github.com/twentyhq) / [twenty](https://github.com/twentyhq/twenty)                                              | 24195 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog](https://github.com/PostHog/posthog)                                                | 22618 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/121462774?s=40&v=4" width="20" height="20" alt=""> &nbsp; [BerriAI](https://github.com/BerriAI) / [litellm](https://github.com/BerriAI/litellm)                                               | 15151 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/179202840?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mediar-ai](https://github.com/mediar-ai) / [screenpipe](https://github.com/mediar-ai/screenpipe)                                   | 11037 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/105877416?s=40&v=4" width="20" height="20" alt=""> &nbsp; [formbricks](https://github.com/formbricks) / [formbricks](https://github.com/formbricks/formbricks)                                |  9386 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/76263028?s=40&v=4" width="20" height="20" alt=""> &nbsp; [anthropics](https://github.com/anthropics) / [courses](https://github.com/anthropics/courses)                                       |  8385 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/78410652?s=40&v=4" width="20" height="20" alt=""> &nbsp; [GreyDGL](https://github.com/GreyDGL) / [PentestGPT](https://github.com/GreyDGL/PentestGPT)                                          |  7374 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/152537519?s=40&v=4" width="20" height="20" alt=""> &nbsp; [superagent-ai](https://github.com/superagent-ai) / [superagent](https://github.com/superagent-ai/superagent)                       |  5391 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/137907881?s=40&v=4" width="20" height="20" alt=""> &nbsp; [promptfoo](https://github.com/promptfoo) / [promptfoo](https://github.com/promptfoo/promptfoo)                                     |  4976 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/157326433?s=40&v=4" width="20" height="20" alt=""> &nbsp; [onlook-dev](https://github.com/onlook-dev) / [onlook](https://github.com/onlook-dev/onlook)                                        |  4141 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/7250217?s=40&v=4" width="20" height="20" alt=""> &nbsp; [Canner](https://github.com/Canner) / [WrenAI](https://github.com/Canner/WrenAI)                                                      |  2526 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/11855343?s=40&v=4" width="20" height="20" alt=""> &nbsp; [pingcap](https://github.com/pingcap) / [autoflow](https://github.com/pingcap/autoflow)                                              |  2061 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85268109?s=40&v=4" width="20" height="20" alt=""> &nbsp; [MLSysOps](https://github.com/MLSysOps) / [MLE-agent](https://github.com/MLSysOps/MLE-agent)                                         |  1161 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [pipelines](https://github.com/open-webui/pipelines)                                  |  1100 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/18422723?s=40&v=4" width="20" height="20" alt=""> &nbsp; [alishobeiri](https://github.com/alishobeiri) / [thread](https://github.com/alishobeiri/thread)                                      |  1074 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/125468716?s=40&v=4" width="20" height="20" alt=""> &nbsp; [topoteretes](https://github.com/topoteretes) / [cognee](https://github.com/topoteretes/cognee)                                     |   971 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/188657705?s=40&v=4" width="20" height="20" alt=""> &nbsp; [bRAGAI](https://github.com/bRAGAI) / [bRAG-langchain](https://github.com/bRAGAI/bRAG-langchain)                                    |   823 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169500408?s=40&v=4" width="20" height="20" alt=""> &nbsp; [opslane](https://github.com/opslane) / [opslane](https://github.com/opslane/opslane)                                               |   677 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/151867818?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dynamiq-ai](https://github.com/dynamiq-ai) / [dynamiq](https://github.com/dynamiq-ai/dynamiq)                                      |   639 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/48585267?s=40&v=4" width="20" height="20" alt=""> &nbsp; [theopenconversationkit](https://github.com/theopenconversationkit) / [tock](https://github.com/theopenconversationkit/tock)         |   514 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/20493493?s=40&v=4" width="20" height="20" alt=""> &nbsp; [andysingal](https://github.com/andysingal) / [llm-course](https://github.com/andysingal/llm-course)                                 |   394 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/132396805?s=40&v=4" width="20" height="20" alt=""> &nbsp; [phospho-app](https://github.com/phospho-app) / [phospho](https://github.com/phospho-app/phospho)                                   |   384 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/178644984?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sentient-engineering](https://github.com/sentient-engineering) / [agent-q](https://github.com/sentient-engineering/agent-q)        |   370 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/168552753?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sql-agi](https://github.com/sql-agi) / [DB-GPT](https://github.com/sql-agi/DB-GPT)                                                 |   324 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog-foss](https://github.com/PostHog/posthog-foss)                                      |   305 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/154247157?s=40&v=4" width="20" height="20" alt=""> &nbsp; [vespperhq](https://github.com/vespperhq) / [vespper](https://github.com/vespperhq/vespper)                                         |   304 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/185116535?s=40&v=4" width="20" height="20" alt=""> &nbsp; [block](https://github.com/block) / [goose](https://github.com/block/goose)                                                         |   295 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/609489?s=40&v=4" width="20" height="20" alt=""> &nbsp; [aorwall](https://github.com/aorwall) / [moatless-tools](https://github.com/aorwall/moatless-tools)                                    |   291 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/2357342?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dmayboroda](https://github.com/dmayboroda) / [minima](https://github.com/dmayboroda/minima)                                          |   221 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/66303003?s=40&v=4" width="20" height="20" alt=""> &nbsp; [RobotecAI](https://github.com/RobotecAI) / [rai](https://github.com/RobotecAI/rai)                                                  |   172 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/148684274?s=40&v=4" width="20" height="20" alt=""> &nbsp; [i-am-alice](https://github.com/i-am-alice) / [3rd-devs](https://github.com/i-am-alice/3rd-devs)                                    |   148 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/171735272?s=40&v=4" width="20" height="20" alt=""> &nbsp; [8090-inc](https://github.com/8090-inc) / [xrx-sample-apps](https://github.com/8090-inc/xrx-sample-apps)                            |   138 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/104478511?s=40&v=4" width="20" height="20" alt=""> &nbsp; [babelcloud](https://github.com/babelcloud) / [LLM-RGB](https://github.com/babelcloud/LLM-RGB)                                      |   135 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/15125613?s=40&v=4" width="20" height="20" alt=""> &nbsp; [souzatharsis](https://github.com/souzatharsis) / [tamingLLMs](https://github.com/souzatharsis/tamingLLMs)                           |   129 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169401942?s=40&v=4" width="20" height="20" alt=""> &nbsp; [LibreChat-AI](https://github.com/LibreChat-AI) / [librechat.ai](https://github.com/LibreChat-AI/librechat.ai)                      |   128 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/51827949?s=40&v=4" width="20" height="20" alt=""> &nbsp; [deepset-ai](https://github.com/deepset-ai) / [haystack-core-integrations](https://github.com/deepset-ai/haystack-core-integrations) |   126 |
+
+## 🔒 セキュリティとプライバシー
+
+データのセキュリティとプライバシーは非常に重要です。  
+詳細につきましては、[セキュリティとプライバシー](https://langfuse.com/security)ページをご参照ください。
+
+### テレメトリー
+
+デフォルトでは、Langfuseはセルフホストされたインスタンスの基本的な使用統計情報を中央サーバ（PostHog）へ自動的に報告します。
+
+これにより、以下が可能となります:
+
+1. Langfuseの利用状況を把握し、最も重要な機能の改善に役立てる
+2. 内部および外部（例：資金調達）のレポートのために全体の利用状況を追跡する
+
+収集されたデータは第三者と共有されず、機微な情報は一切含まれていません。  
+当社はこの点について極力透明性を保っており、収集される具体的なデータの詳細は[こちら](/web/src/features/telemetry/index.ts)で確認できます。
+
+`TELEMETRY_ENABLED=false` を設定することで、テレメトリーの報告をオプトアウトできます.

README.kr.md

@@ -0,0 +1,339 @@
+![Langfuse GitHub Banner](https://langfuse.com/images/docs/github-readme/github-banner.png)
+
+<div align="center">
+   <div>
+      <h3>
+      <a href="https://langfuse.com/kr">
+            <strong>🇰🇷 🤝 🪢</strong>
+         </a> · 
+         <a href="https://cloud.langfuse.com">
+            <strong>Langfuse Cloud</strong>
+         </a> · 
+         <a href="https://langfuse.com/docs/deployment/self-host">
+            <strong>셀프 호스트</strong>
+         </a> · 
+         <a href="https://langfuse.com/demo">
+            <strong>데모</strong>
+         </a>
+      </h3>
+   </div>
+
+   <div>
+      <a href="https://langfuse.com/docs"><strong>문서</strong></a> ·
+      <a href="https://langfuse.com/issues"><strong>버그 신고</strong></a> ·
+      <a href="https://langfuse.com/ideas"><strong>기능 요청</strong></a> ·
+      <a href="https://langfuse.com/changelog"><strong>변경 내역</strong></a> ·
+      <a href="https://langfuse.com/roadmap"><strong>로드맵</strong></a> ·
+   </div>
+   <br/>
+   <span>Langfuse는 지원 및 기능 요청을 위해 <a href="https://github.com/orgs/langfuse/discussions"><strong>GitHub Discussions</strong></a>를 사용합니다.</span>
+   <br/>
+   <span><b>채용 중입니다.</b> <a href="https://langfuse.com/careers"><strong>함께 하세요</strong></a> – 제품 엔지니어링 및 기술 go-to-market 역할의 인재를 찾고 있습니다.</span>
+   <br/>
+   <br/>
+   <div>
+   </div>
+</div>
+
+<p align="center">
+   <a href="https://github.com/langfuse/langfuse/blob/main/LICENSE">
+   <img src="https://img.shields.io/badge/License-MIT-E11311.svg" alt="MIT License">
+   </a>
+   <a href="https://www.ycombinator.com/companies/langfuse"><img src="https://img.shields.io/badge/Y%20Combinator-W23-orange" alt="Y Combinator W23"></a>
+   <a href="https://hub.docker.com/u/langfuse" target="_blank">
+   <img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langfuse/langfuse?labelColor=%20%23FDB062&logo=Docker&labelColor=%20%23528bff"></a>
+   <a href="https://pypi.python.org/pypi/langfuse"><img src="https://img.shields.io/pypi/dm/langfuse?logo=python&logoColor=white&label=pypi%20langfuse&color=blue" alt="langfuse Python package on PyPi"></a>
+   <a href="https://www.npmjs.com/package/langfuse"><img src="https://img.shields.io/npm/dm/langfuse?logo=npm&logoColor=white&label=npm%20langfuse&color=blue" alt="langfuse npm package"></a>
+   <br/>
+   <a href="https://discord.com/invite/7NXusRtqYU" target="_blank">
+   <img src="https://img.shields.io/discord/1111061815649124414?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
+      alt="chat on Discord"></a>
+   <a href="https://twitter.com/intent/follow?screen_name=langfuse" target="_blank">
+   <img src="https://img.shields.io/twitter/follow/langfuse?logo=X&color=%20%23f5f5f5"
+      alt="follow on X(Twitter)"></a>
+   <a href="https://www.linkedin.com/company/langfuse/" target="_blank">
+   <img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
+      alt="follow on LinkedIn"></a>
+   <a href="https://github.com/langfuse/langfuse/graphs/commit-activity" target="_blank">
+   <img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langfuse/langfuse?labelColor=%20%2332b583&color=%20%2312b76a"></a>
+   <a href="https://github.com/langfuse/langfuse/" target="_blank">
+   <img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alangfuse%2Flangfuse%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
+   <a href="https://github.com/langfuse/langfuse/discussions/" target="_blank">
+   <img alt="Discussion posts" src="https://img.shields.io/github/discussions/langfuse/langfuse?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
+</p>
+
+<p align="center">
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README.cn.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README.ja.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README.kr.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+</p>
+
+Langfuse는 **오픈 소스 LLM 엔지니어링** 플랫폼입니다.  
+팀이 협업하여 AI 애플리케이션을 **개발, 모니터링, 평가** 및 **디버그**할 수 있도록 도와줍니다.  
+Langfuse는 몇 분 안에 셀프 호스팅할 수 있으며, 검증된(battle-tested) 솔루션입니다.
+
+[![Langfuse Overview Video](https://github.com/user-attachments/assets/3926b288-ff61-4b95-8aa1-45d041c70866)](https://langfuse.com/watch-demo)
+
+## ✨ 주요 기능
+
+![Langfuse Overview](https://langfuse.com/images/docs/github-readme/github-feature-overview.png)
+
+- **LLM 애플리케이션 관측**  
+  앱에 계측(instrumentation)을 추가하여 Langfuse로 trace 데이터를 수집함으로써, 검색, 임베딩, 또는 에이전트 동작과 같은 LLM 호출 및 기타 관련 로직을 추적할 수 있습니다. 복잡한 로그와 사용자 세션을 확인 및 디버깅 해보세요. 인터랙티브 데모를 통해 실제 작동 예를 확인할 수 있습니다.
+
+- **프롬프트 관리**  
+  프롬프트를 중앙에서 관리하고 버전 관리하며 협업으로 수정할 수 있도록 도와줍니다. 서버와 클라이언트 측의 강력한 캐싱 덕분에 애플리케이션에 지연(latency)을 추가하지 않고도 프롬프트를 반복 개선할 수 있습니다.
+
+- **평가**  
+  LLM 애플리케이션 개발 워크플로우에서 핵심적인 역할을 하며, Langfuse는 여러분의 필요에 맞게 유연하게 대응합니다. LLM을 심사자로 활용하는 기능, 사용자 피드백 수집, 수동 라벨링 및 API/SDK를 통한 맞춤 평가 파이프라인을 지원합니다.
+
+- **데이터셋**  
+  LLM 애플리케이션 평가를 위한 테스트 세트와 벤치마크를 제공하여, 지속적인 개선, 배포 전 테스트, 구조화된 실험, 유연한 평가 및 LangChain과 LlamaIndex와 같은 프레임워크와의 원활한 통합을 지원합니다.
+
+- **LLM 플레이그라운드**  
+  프롬프트와 모델 구성에 대해 테스트 및 반복 개선할 수 있는 도구로, 피드백 루프를 단축하여 개발 속도를 높여줍니다. trace에서 이상한 결과가 발생하면 플레이그라운드로 바로 이동해 개선할 수 있습니다.
+
+- **종합 API**  
+  Langfuse는 API를 통해 제공되는 구성 요소들을 활용하여 맞춤형 LLMOps 워크플로우를 강화하는 데 자주 사용됩니다. OpenAPI 명세, Postman 컬렉션, Python 및 JS/TS용 타입드 SDK가 제공됩니다.
+
+## 📦 Langfuse 배포
+
+![Langfuse Deployment Options](https://langfuse.com/images/docs/github-readme/github-deployment-options.png)
+
+### Langfuse Cloud
+
+Langfuse 팀이 관리하는 배포 방식으로, 후한 무료 플랜(취미 플랜)을 제공하며 신용카드가 필요하지 않습니다.
+
+<div align="center">
+    <a href="https://cloud.langfuse.com" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/»%20Sign%20up%20for%20Langfuse%20Cloud-8A2BE2?&color=orange">
+    </a>
+</div>
+
+### Langfuse 셀프 호스트
+
+자체 인프라에서 Langfuse를 실행하세요:
+
+- [로컬 (docker compose)](https://langfuse.com/self-hosting/local): Docker Compose를 사용하여 본인의 컴퓨터에서 5분 안에 Langfuse를 실행할 수 있습니다.
+
+  ```bash
+  # 최신 Langfuse 저장소 클론
+  git clone https://github.com/langfuse/langfuse.git
+  cd langfuse
+
+  # langfuse docker compose 실행
+  docker compose up
+  ```
+
+- [Kubernetes (Helm)](https://langfuse.com/self-hosting/kubernetes-helm): Helm을 사용해 Kubernetes 클러스터에서 Langfuse를 실행합니다. 이는 권장되는 프로덕션 배포 방식입니다.
+- [VM](https://langfuse.com/self-hosting/docker-compose): Docker Compose를 사용해 단일 가상 머신에서 Langfuse를 실행합니다.
+- 예정: 클라우드별 배포 가이드 – 아래 스레드에서 투표 및 댓글을 남겨주세요: [AWS](https://github.com/orgs/langfuse/discussions/4645), [Google Cloud](https://github.com/orgs/langfuse/discussions/4646), [Azure](https://github.com/orgs/langfuse/discussions/4647).
+
+자세한 내용은 [자체 호스팅 문서](https://langfuse.com/self-hosting)를 참조하세요.
+
+## 🔌 통합 기능
+
+![Langfuse Integrations](https://langfuse.com/images/docs/github-readme/github-integrations.png)
+
+### 주요 통합:
+
+| 통합                                                                 | 지원                   | 설명                                                                                                                                      |
+| --------------------------------------------------------------------- | ---------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
+| [SDK](https://langfuse.com/docs/sdk)                                  | Python, JS/TS          | SDK를 사용하여 완전한 유연성을 갖춘 수동 계측(manual instrumentation)을 수행합니다.                                                       |
+| [OpenAI](https://langfuse.com/docs/integrations/openai)               | Python, JS/TS          | OpenAI SDK의 드롭인 대체(drop-in replacement)를 통해 자동 계측(automated instrumentation)을 수행합니다.                                       |
+| [Langchain](https://langfuse.com/docs/integrations/langchain)         | Python, JS/TS          | Langchain 애플리케이션에 callback 핸들러를 전달하여 자동 계측합니다.                                                                       |
+| [LlamaIndex](https://langfuse.com/docs/integrations/llama-index/get-started) | Python                 | LlamaIndex 콜백 시스템을 통한 자동 계측을 지원합니다.                                                                                      |
+| [Haystack](https://langfuse.com/docs/integrations/haystack)           | Python                 | Haystack 콘텐츠 추적 시스템을 통한 자동 계측을 지원합니다.                                                                                 |
+| [LiteLLM](https://langfuse.com/docs/integrations/litellm)             | Python, JS/TS (proxy only) | GPT의 드롭인 대체품으로 어떤 LLM도 사용할 수 있습니다. Azure, OpenAI, Cohere, Anthropic, Ollama, VLLM, Sagemaker, HuggingFace, Replicate 등 100개 이상의 LLM 지원. |
+| [Vercel AI SDK](https://langfuse.com/docs/integrations/vercel-ai-sdk)   | JS/TS                  | React, Next.js, Vue, Svelte, Node.js와 함께 AI 기반 애플리케이션 구축을 돕는 TypeScript 툴킷입니다.                                             |
+| [API](https://langfuse.com/docs/api)                                   |                        | 공개 API를 직접 호출합니다. OpenAPI 명세가 제공됩니다.                                                                                      |
+
+### Langfuse와 통합된 패키지:
+
+| 이름                                                                    | 유형               | 설명                                                                                                             |
+| ----------------------------------------------------------------------- | ------------------ | ---------------------------------------------------------------------------------------------------------------- |
+| [Instructor](https://langfuse.com/docs/integrations/instructor)         | 라이브러리         | 구조화된 LLM 출력을(JSON, Pydantic) 얻기 위한 라이브러리입니다.                                                  |
+| [DSPy](https://langfuse.com/docs/integrations/dspy)                     | 라이브러리         | 언어 모델 프롬프트와 가중치를 체계적으로 최적화하는 프레임워크입니다.                                             |
+| [Mirascope](https://langfuse.com/docs/integrations/mirascope)           | 라이브러리         | LLM 애플리케이션 구축을 위한 Python 툴킷입니다.                                                                    |
+| [Ollama](https://langfuse.com/docs/integrations/ollama)                 | 모델 (로컬)        | 자신의 컴퓨터에서 오픈 소스 LLM을 손쉽게 실행할 수 있습니다.                                                      |
+| [Amazon Bedrock](https://langfuse.com/docs/integrations/amazon-bedrock) | 모델               | AWS에서 기본 및 파인튜닝된 모델을 실행합니다.                                                                      |
+| [AutoGen](https://langfuse.com/docs/integrations/autogen)               | 에이전트 프레임워크 | 분산 에이전트 구축을 위한 오픈 소스 LLM 플랫폼입니다.                                                              |
+| [Flowise](https://langfuse.com/docs/integrations/flowise)               | 채팅/에이전트 UI    | 맞춤형 LLM 플로우를 위한 JS/TS 코드 없는(no-code) 빌더입니다.                                                       |
+| [Langflow](https://langfuse.com/docs/integrations/langflow)             | 채팅/에이전트 UI    | react-flow를 활용하여 실험 및 프로토타이핑을 손쉽게 할 수 있도록 디자인된 LangChain용 Python 기반 UI입니다.          |
+| [Dify](https://langfuse.com/docs/integrations/dify)                     | 채팅/에이전트 UI    | 코드 없는 빌더와 함께 제공되는 오픈 소스 LLM 애플리케이션 개발 플랫폼입니다.                                        |
+| [OpenWebUI](https://langfuse.com/docs/integrations/openwebui)           | 채팅/에이전트 UI    | 셀프 호스팅 및 로컬 모델 등 다양한 LLM 실행기를 지원하는 셀프 호스팅 LLM 채팅 웹 UI입니다.                              |
+| [Promptfoo](https://langfuse.com/docs/integrations/promptfoo)           | 도구               | 오픈 소스 LLM 테스트 플랫폼입니다.                                                                                 |
+| [LobeChat](https://langfuse.com/docs/integrations/lobechat)             | 채팅/에이전트 UI    | 오픈 소스 챗봇 플랫폼입니다.                                                                                         |
+| [Vapi](https://langfuse.com/docs/integrations/vapi)                     | 플랫폼             | 오픈 소스 음성 AI 플랫폼입니다.                                                                                      |
+| [Inferable](https://langfuse.com/docs/integrations/other/inferable)     | 에이전트           | 분산 에이전트 구축을 위한 오픈 소스 LLM 플랫폼입니다.                                                              |
+| [Gradio](https://langfuse.com/docs/integrations/other/gradio)           | 채팅/에이전트 UI    | 채팅 UI와 같은 웹 인터페이스 구축을 위한 오픈 소스 Python 라이브러리입니다.                                         |
+| [Goose](https://langfuse.com/docs/integrations/goose)                   | 에이전트           | 분산 에이전트 구축을 위한 오픈 소스 LLM 플랫폼입니다.                                                              |
+| [smolagents](https://langfuse.com/docs/integrations/smolagents)         | 에이전트           | 오픈 소스 AI 에이전트 프레임워크입니다.                                                                            |
+| [CrewAI](https://langfuse.com/docs/integrations/crewai)                 | 에이전트           | 에이전트 간 협업 및 도구 사용을 위한 다중 에이전트 프레임워크입니다.                                                 |
+
+## 🚀 빠른 시작
+
+앱에 계측을 추가하고 Langfuse에 trace 데이터를 수집하여, LLM 호출 및 검색, 임베딩, 에이전트 동작과 같은 애플리케이션 로직을 추적해보세요. 복잡한 로그와 사용자 세션을 확인하여 디버깅할 수 있습니다.
+
+### 1️⃣ 새 프로젝트 생성
+
+1. [Langfuse 계정 생성](https://cloud.langfuse.com/auth/sign-up) 또는 [셀프 호스트](https://langfuse.com/self-hosting)
+2. 새 프로젝트를 생성합니다.
+3. 프로젝트 설정에서 새로운 API 자격 증명을 생성합니다.
+
+### 2️⃣ 첫 번째 LLM 호출 기록하기
+
+[`@observe()` 데코레이터](https://langfuse.com/docs/sdk/python/decorators)를 사용하면 Python LLM 애플리케이션의 추적이 매우 간편해집니다. 이 빠른 시작 예제에서는 Langfuse [OpenAI 통합](https://langfuse.com/docs/integrations/openai)을 사용하여 모든 모델 파라미터를 자동으로 캡처합니다.
+
+> [!TIP]
+> OpenAI를 사용하지 않으시다면, 다른 모델 및 프레임워크의 로그 기록 방법은 [문서](https://langfuse.com/docs/get-started#log-your-first-llm-call-to-langfuse)를 참조하세요.
+
+```bash
+pip install langfuse openai
+```
+
+```bash filename=".env"
+LANGFUSE_SECRET_KEY="sk-lf-..."
+LANGFUSE_PUBLIC_KEY="pk-lf-..."
+LANGFUSE_HOST="https://cloud.langfuse.com" # 🇪🇺 EU region
+# LANGFUSE_HOST="https://us.cloud.langfuse.com" # 🇺🇸 US region
+```
+
+```python:main.py
+from langfuse.decorators import observe
+from langfuse.openai import openai # OpenAI integration
+
+@observe()
+def story():
+    return openai.chat.completions.create(
+        model="gpt-4o",
+        messages=[{"role": "user", "content": "What is Langfuse?"}],
+    ).choices[0].message.content
+
+@observe()
+def main():
+    return story()
+
+main()
+```
+
+### 3️⃣ Langfuse에서 trace 확인하기
+
+Langfuse에서 LLM 호출 및 애플리케이션의 기타 로직에 대한 trace를 확인할 수 있습니다.
+
+![Example trace in Langfuse](https://langfuse.com/images/docs/github-readme/github-example-trace.png)
+
+_[Langfuse의 공개 예제 trace](https://cloud.langfuse.com/project/cloramnkj0002jz088vzn1ja4/traces/2cec01e3-3dc2-472f-afcf-3b968cf0c1f4?timestamp=2025-02-10T14%3A27%3A30.275Z&observation=cb5ff844-07ef-41e6-b8e2-6c64344bc13b)_
+
+> [!TIP]
+>
+> Langfuse의 trace에 대해 더 알아보거나 [인터랙티브 데모](https://langfuse.com/docs/demo)에서 직접 체험해보세요.
+
+## ⭐️ 별을 눌러주세요
+
+![Star Langfuse](https://langfuse.com/images/docs/github-readme/github-star-howto.gif)
+
+## 💭 지원
+
+질문에 대한 답변을 찾는 방법:
+
+- 우리의 [문서](https://langfuse.com/docs)는 답을 찾기 위한 최적의 장소입니다. 문서가 매우 포괄적이며, 유지보수에 많은 노력을 기울이고 있습니다. GitHub를 통해 문서 수정 제안도 가능합니다.
+- [Langfuse FAQ](https://langfuse.com/faq)에서는 가장 흔한 질문에 대해 답변하고 있습니다.
+- 질문에 즉각적인 답변이 필요하다면 [Ask AI](https://langfuse.com/docs/ask-ai)를 사용해보세요.
+
+지원 채널:
+
+- **GitHub Discussions의 [공개 Q&A](https://github.com/orgs/langfuse/discussions/categories/support)** 에 질문을 남겨주세요. 가능한 한 많은 세부 사항(예: 코드 스니펫, 스크린샷, 배경 정보)을 포함해 질문해 주시기 바랍니다.
+- [기능 요청](https://github.com/orgs/langfuse/discussions/categories/ideas)을 남겨주세요.
+- [버그 신고](https://github.com/langfuse/langfuse/issues)는 GitHub Issues를 통해 해주세요.
+- 긴급한 문의는 앱 내 채팅 위젯을 통해 연락 바랍니다.
+
+## 🤝 기여하기
+
+여러분의 기여를 환영합니다!
+
+- GitHub Discussions의 [아이디어](https://github.com/orgs/langfuse/discussions/categories/ideas)에 투표해보세요.
+- GitHub Issues에서 이슈를 제기하고 댓글을 남겨주세요.
+- PR을 제출하세요 – 개발 환경 설정 방법 등 자세한 내용은 [CONTRIBUTING.md](CONTRIBUTING.md)를 참조하세요.
+
+## 🥇 라이선스
+
+이 저장소는 `ee` 폴더를 제외하고 MIT 라이선스가 적용됩니다. 자세한 내용은 [LICENSE](LICENSE)와 [문서](https://langfuse.com/docs/open-source)를 확인하세요.
+
+## ⭐️ 별(Star) 히스토리
+
+<a href="https://star-history.com/#langfuse/langfuse&Date">
+ <picture>
+   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date&theme=dark" />
+   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
+   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" style="border-radius: 15px;" />
+ </picture>
+</a>
+
+## ❤️ Langfuse를 사용하는 오픈 소스 프로젝트
+
+별(star) 수를 기준으로 순위가 매겨진 Langfuse를 사용하는 상위 오픈 소스 Python 프로젝트들 ([출처](https://github.com/langfuse/langfuse-docs/blob/main/components-mdx/dependents)):
+
+| 저장소                                                                                                                                                                                                                                                          | 별   |
+| :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---: |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/127165244?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langgenius](https://github.com/langgenius) / [dify](https://github.com/langgenius/dify)                                            | 54865 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [open-webui](https://github.com/open-webui/open-webui)                                | 51531 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/131470832?s=40&v=4" width="20" height="20" alt=""> &nbsp; [lobehub](https://github.com/lobehub) / [lobe-chat](https://github.com/lobehub/lobe-chat)                                           | 49003 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85702467?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langflow-ai](https://github.com/langflow-ai) / [langflow](https://github.com/langflow-ai/langflow)                                  | 39093 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/130722866?s=40&v=4" width="20" height="20" alt=""> &nbsp; [run-llama](https://github.com/run-llama) / [llama_index](https://github.com/run-llama/llama_index)                                 | 37368 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/139558948?s=40&v=4" width="20" height="20" alt=""> &nbsp; [chatchat-space](https://github.com/chatchat-space) / [Langchain-Chatchat](https://github.com/chatchat-space/Langchain-Chatchat)    | 32486 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/128289781?s=40&v=4" width="20" height="20" alt=""> &nbsp; [FlowiseAI](https://github.com/FlowiseAI) / [Flowise](https://github.com/FlowiseAI/Flowise)                                         | 32448 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/31035808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mindsdb](https://github.com/mindsdb) / [mindsdb](https://github.com/mindsdb/mindsdb)                                                | 26931 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/119600397?s=40&v=4" width="20" height="20" alt=""> &nbsp; [twentyhq](https://github.com/twentyhq) / [twenty](https://github.com/twentyhq/twenty)                                              | 24195 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog](https://github.com/PostHog/posthog)                                                | 22618 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/121462774?s=40&v=4" width="20" height="20" alt=""> &nbsp; [BerriAI](https://github.com/BerriAI) / [litellm](https://github.com/BerriAI/litellm)                                               | 15151 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/179202840?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mediar-ai](https://github.com/mediar-ai) / [screenpipe](https://github.com/mediar-ai/screenpipe)                                   | 11037 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/105877416?s=40&v=4" width="20" height="20" alt=""> &nbsp; [formbricks](https://github.com/formbricks) / [formbricks](https://github.com/formbricks/formbricks)                                |  9386 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/76263028?s=40&v=4" width="20" height="20" alt=""> &nbsp; [anthropics](https://github.com/anthropics) / [courses](https://github.com/anthropics/courses)                                       |  8385 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/78410652?s=40&v=4" width="20" height="20" alt=""> &nbsp; [GreyDGL](https://github.com/GreyDGL) / [PentestGPT](https://github.com/GreyDGL/PentestGPT)                                          |  7374 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/152537519?s=40&v=4" width="20" height="20" alt=""> &nbsp; [superagent-ai](https://github.com/superagent-ai) / [superagent](https://github.com/superagent-ai/superagent)                       |  5391 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/137907881?s=40&v=4" width="20" height="20" alt=""> &nbsp; [promptfoo](https://github.com/promptfoo) / [promptfoo](https://github.com/promptfoo/promptfoo)                                     |  4976 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/157326433?s=40&v=4" width="20" height="20" alt=""> &nbsp; [onlook-dev](https://github.com/onlook-dev) / [onlook](https://github.com/onlook-dev/onlook)                                        |  4141 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/7250217?s=40&v=4" width="20" height="20" alt=""> &nbsp; [Canner](https://github.com/Canner) / [WrenAI](https://github.com/Canner/WrenAI)                                                      |  2526 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/11855343?s=40&v=4" width="20" height="20" alt=""> &nbsp; [pingcap](https://github.com/pingcap) / [autoflow](https://github.com/pingcap/autoflow)                                              |  2061 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85268109?s=40&v=4" width="20" height="20" alt=""> &nbsp; [MLSysOps](https://github.com/MLSysOps) / [MLE-agent](https://github.com/MLSysOps/MLE-agent)                                         |  1161 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [pipelines](https://github.com/open-webui/pipelines)                                  |  1100 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/18422723?s=40&v=4" width="20" height="20" alt=""> &nbsp; [alishobeiri](https://github.com/alishobeiri) / [thread](https://github.com/alishobeiri/thread)                                      |  1074 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/125468716?s=40&v=4" width="20" height="20" alt=""> &nbsp; [topoteretes](https://github.com/topoteretes) / [cognee](https://github.com/topoteretes/cognee)                                     |   971 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/188657705?s=40&v=4" width="20" height="20" alt=""> &nbsp; [bRAGAI](https://github.com/bRAGAI) / [bRAG-langchain](https://github.com/bRAGAI/bRAG-langchain)                                    |   823 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169500408?s=40&v=4" width="20" height="20" alt=""> &nbsp; [opslane](https://github.com/opslane) / [opslane](https://github.com/opslane/opslane)                                               |   677 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/151867818?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dynamiq-ai](https://github.com/dynamiq-ai) / [dynamiq](https://github.com/dynamiq-ai/dynamiq)                                      |   639 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/48585267?s=40&v=4" width="20" height="20" alt=""> &nbsp; [theopenconversationkit](https://github.com/theopenconversationkit) / [tock](https://github.com/theopenconversationkit/tock)         |   514 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/20493493?s=40&v=4" width="20" height="20" alt=""> &nbsp; [andysingal](https://github.com/andysingal) / [llm-course](https://github.com/andysingal/llm-course)                                 |   394 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/132396805?s=40&v=4" width="20" height="20" alt=""> &nbsp; [phospho-app](https://github.com/phospho-app) / [phospho](https://github.com/phospho-app/phospho)                                   |   384 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/178644984?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sentient-engineering](https://github.com/sentient-engineering) / [agent-q](https://github.com/sentient-engineering/agent-q)        |   370 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/168552753?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sql-agi](https://github.com/sql-agi) / [DB-GPT](https://github.com/sql-agi/DB-GPT)                                                 |   324 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog-foss](https://github.com/PostHog/posthog-foss)                                      |   305 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/154247157?s=40&v=4" width="20" height="20" alt=""> &nbsp; [vespperhq](https://github.com/vespperhq) / [vespper](https://github.com/vespperhq/vespper)                                         |   304 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/185116535?s=40&v=4" width="20" height="20" alt=""> &nbsp; [block](https://github.com/block) / [goose](https://github.com/block/goose)                                                         |   295 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/609489?s=40&v=4" width="20" height="20" alt=""> &nbsp; [aorwall](https://github.com/aorwall) / [moatless-tools](https://github.com/aorwall/moatless-tools)                                    |   291 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/2357342?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dmayboroda](https://github.com/dmayboroda) / [minima](https://github.com/dmayboroda/minima)                                          |   221 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/66303003?s=40&v=4" width="20" height="20" alt=""> &nbsp; [RobotecAI](https://github.com/RobotecAI) / [rai](https://github.com/RobotecAI/rai)                                                  |   172 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/148684274?s=40&v=4" width="20" height="20" alt=""> &nbsp; [i-am-alice](https://github.com/i-am-alice) / [3rd-devs](https://github.com/i-am-alice/3rd-devs)                                    |   148 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/171735272?s=40&v=4" width="20" height="20" alt=""> &nbsp; [8090-inc](https://github.com/8090-inc) / [xrx-sample-apps](https://github.com/8090-inc/xrx-sample-apps)                            |   138 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/104478511?s=40&v=4" width="20" height="20" alt=""> &nbsp; [babelcloud](https://github.com/babelcloud) / [LLM-RGB](https://github.com/babelcloud/LLM-RGB)                                      |   135 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/15125613?s=40&v=4" width="20" height="20" alt=""> &nbsp; [souzatharsis](https://github.com/souzatharsis) / [tamingLLMs](https://github.com/souzatharsis/tamingLLMs)                           |   129 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169401942?s=40&v=4" width="20" height="20" alt=""> &nbsp; [LibreChat-AI](https://github.com/LibreChat-AI) / [librechat.ai](https://github.com/LibreChat-AI/librechat.ai)                      |   128 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/51827949?s=40&v=4" width="20" height="20" alt=""> &nbsp; [deepset-ai](https://github.com/deepset-ai) / [haystack-core-integrations](https://github.com/deepset-ai/haystack-core-integrations) |   126 |
+
+## 🔒 보안 & 개인정보 보호
+
+우리는 데이터 보안과 개인정보 보호를 매우 중요하게 생각합니다. 자세한 내용은 [Security and Privacy](https://langfuse.com/security) 페이지를 참조하세요.
+
+### 텔레메트리
+
+기본적으로 Langfuse는 자체 호스팅 인스턴스의 기본 사용 통계를 중앙 서버(PostHog)로 자동 보고합니다.
+
+이를 통해:
+1. Langfuse가 어떻게 사용되는지 이해하고, 가장 관련성 높은 기능을 개선할 수 있습니다.
+2. 내부 및 외부(예: 자금 조달) 보고를 위한 전체 사용량을 추적할 수 있습니다.
+
+수집된 데이터는 제3자와 공유되지 않으며 민감한 정보를 포함하지 않습니다. 이에 대해 매우 투명하게 공개하고 있으며, 수집되는 정확한 데이터는 [여기](/web/src/features/telemetry/index.ts)에서 확인할 수 있습니다.
+
+환경 변수 `TELEMETRY_ENABLED=false`를 설정하여 옵트아웃할 수 있습니다.

README.md

@@ -1,115 +1,132 @@
-![Langfuse GitHub Banner](https://github.com/langfuse/langfuse/assets/121163007/6035f0f3-d691-4963-b5d0-10cf506e9d42)
-
-<div align="center"><h1>Langfuse: Open Source LLM Engineering Platform</h1></div>
-<div align="center"><h4>LLM Observability, Prompt Management, LLM Evaluations,<br/>Datasets, LLM Metrics, and Prompt Playground</h4></div>
+![Langfuse GitHub Banner](https://langfuse.com/images/docs/github-readme/github-banner.png)
 
 <div align="center">
    <div>
       <h3>
          <a href="https://cloud.langfuse.com">
-            <strong>Sign up</strong>
+            <strong>Langfuse Cloud</strong>
          </a> · 
          <a href="https://langfuse.com/docs/deployment/self-host">
             <strong>Self Host</strong>
          </a> · 
          <a href="https://langfuse.com/demo">
-            <strong>Demo (live data)</strong>
+            <strong>Demo</strong>
          </a>
       </h3>
    </div>
+
    <div>
       <a href="https://langfuse.com/docs"><strong>Docs</strong></a> ·
       <a href="https://langfuse.com/issues"><strong>Report Bug</strong></a> ·
       <a href="https://langfuse.com/ideas"><strong>Feature Request</strong></a> ·
       <a href="https://langfuse.com/changelog"><strong>Changelog</strong></a> ·
       <a href="https://langfuse.com/roadmap"><strong>Roadmap</strong></a> ·
-      <a href="https://langfuse.com/discord"><strong>Discord</strong></a> 
    </div>
+   <br/>
    <span>Langfuse uses <a href="https://github.com/orgs/langfuse/discussions"><strong>Github Discussions</strong></a>  for Support and Feature Requests.</span>
    <br/>
-   <span>We're hiring. <a href="https://langfuse.com/careers"><strong>Join us</strong></a> in Product Engineering and Developer Relations.</span>
+   <span><b>We're hiring.</b> <a href="https://langfuse.com/careers"><strong>Join us</strong></a> in product engineering and technical go-to-market roles.</span>
    <br/>
    <br/>
    <div>
-      <a href="https://github.com/langfuse/langfuse/blob/main/LICENSE"><img src="https://img.shields.io/badge/License-MIT-red.svg?style=flat-square" alt="MIT License"></a>
-      <a href="https://www.ycombinator.com/companies/langfuse"><img src="https://img.shields.io/badge/Y%20Combinator-W23-orange?style=flat-square" alt="Y Combinator W23"></a>
-      <a href="https://github.com/langfuse/langfuse/pkgs/container/langfuse"><img alt="Docker Image" src="https://img.shields.io/badge/docker-langfuse-blue?logo=Docker&logoColor=white&style=flat-square"></a>
-      <a href="https://pypi.python.org/pypi/langfuse"><img src="https://img.shields.io/pypi/dm/langfuse?style=flat-square&logo=python&logoColor=white&label=pypi%20langfuse&color=blue" alt="langfuse Python package on PyPi"></a>
-      <a href="https://www.npmjs.com/package/langfuse"><img src="https://img.shields.io/npm/dm/langfuse?style=flat-square&logo=npm&logoColor=white&label=npm%20langfuse&color=blue" alt="langfuse npm package"></a>
    </div>
 </div>
-</br>
 
-## Langfuse Overview
+<p align="center">
+   <a href="https://github.com/langfuse/langfuse/blob/main/LICENSE">
+   <img src="https://img.shields.io/badge/License-MIT-E11311.svg" alt="MIT License">
+   </a>
+   <a href="https://www.ycombinator.com/companies/langfuse"><img src="https://img.shields.io/badge/Y%20Combinator-W23-orange" alt="Y Combinator W23"></a>
+   <a href="https://hub.docker.com/u/langfuse" target="_blank">
+   <img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langfuse/langfuse?labelColor=%20%23FDB062&logo=Docker&labelColor=%20%23528bff"></a>
+   <a href="https://pypi.python.org/pypi/langfuse"><img src="https://img.shields.io/pypi/dm/langfuse?logo=python&logoColor=white&label=pypi%20langfuse&color=blue" alt="langfuse Python package on PyPi"></a>
+   <a href="https://www.npmjs.com/package/langfuse"><img src="https://img.shields.io/npm/dm/langfuse?logo=npm&logoColor=white&label=npm%20langfuse&color=blue" alt="langfuse npm package"></a>
+   <br/>
+   <a href="https://discord.com/invite/7NXusRtqYU" target="_blank">
+   <img src="https://img.shields.io/discord/1111061815649124414?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
+      alt="chat on Discord"></a>
+   <a href="https://twitter.com/intent/follow?screen_name=langfuse" target="_blank">
+   <img src="https://img.shields.io/twitter/follow/langfuse?logo=X&color=%20%23f5f5f5"
+      alt="follow on X(Twitter)"></a>
+   <a href="https://www.linkedin.com/company/langfuse/" target="_blank">
+   <img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
+      alt="follow on LinkedIn"></a>
+   <a href="https://github.com/langfuse/langfuse/graphs/commit-activity" target="_blank">
+   <img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langfuse/langfuse?labelColor=%20%2332b583&color=%20%2312b76a"></a>
+   <a href="https://github.com/langfuse/langfuse/" target="_blank">
+   <img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alangfuse%2Flangfuse%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
+   <a href="https://github.com/langfuse/langfuse/discussions/" target="_blank">
+   <img alt="Discussion posts" src="https://img.shields.io/github/discussions/langfuse/langfuse?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
+</p>
 
-_Unmute video for voice-over_
+<p align="center">
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README.cn.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README.ja.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README.kr.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+</p>
 
-https://github.com/langfuse/langfuse/assets/2834609/a94062e9-c782-4ee9-af59-dee6370149a8
+Langfuse is an **open source LLM engineering** platform. It helps teams collaboratively
+**develop, monitor, evaluate,** and **debug** AI applications. Langfuse can be **self-hosted in minutes** and is **battle-tested**.
 
-### Develop
+[![Langfuse Overview Video](https://github.com/user-attachments/assets/3926b288-ff61-4b95-8aa1-45d041c70866)](https://langfuse.com/watch-demo)
 
-- **LLM Observability:** Instrument your app and start ingesting traces to Langfuse ([Quickstart](https://langfuse.com/docs/get-started), [Integrations](https://langfuse.com/docs/integrations) [Tracing](https://langfuse.com/docs/tracing))
-- **Langfuse UI:** Inspect and debug complex logs ([Demo](https://langfuse.com/docs/demo), [Tracing](https://langfuse.com/docs/tracing))
-- **Prompt Management:** Manage, version and deploy prompts from within Langfuse ([Prompt Management](https://langfuse.com/docs/prompts/get-started))
-- **Prompt Engineering:** Test and iterate on your prompts with the [LLM Playground](https://langfuse.com/docs/playground)
+## ✨ Core Features
 
-### Monitor
+![Langfuse Overview](https://langfuse.com/images/docs/github-readme/github-feature-overview.png)
 
-- **LLM Analytics:** Track metrics (cost, latency, quality) and gain insights from dashboards & data exports ([Analytics](https://langfuse.com/docs/analytics))
-- **LLM Evaluations:** Collect and calculate scores for your LLM completions ([Scores & Evaluations](https://langfuse.com/docs/scores))
-  - Run ([Model-based evaluations](https://langfuse.com/docs/scores/model-based-evals)) and LLM-as-a-Judge within Langfuse
-  - Collect user feedback ([User Feedback](https://langfuse.com/docs/scores/user-feedback))
-  - Manually score LLM outputs in Langfuse ([Manual Scores](https://langfuse.com/docs/scores/manually))
+- [LLM Application Observability](https://langfuse.com/docs/tracing): Instrument your app and start ingesting traces to Langfuse, thereby tracking LLM calls and other relevant logic in your app such as retrieval, embedding, or agent actions. Inspect and debug complex logs and user sessions. Try the interactive [demo](https://langfuse.com/docs/demo) to see this in action.
 
-### Test
+- [Prompt Management](https://langfuse.com/docs/prompts/get-started) helps you centrally manage, version control, and collaboratively iterate on your prompts. Thanks to strong caching on server and client side, you can iterate on prompts without adding latency to your application.
 
-- **Experiments:** Track and test app behaviour before deploying a new version
-  - Datasets let you test expected in and output pairs and benchmark performance before deploying ([Datasets](https://langfuse.com/docs/datasets))
-  - Track versions and releases in your application ([Experimentation](https://langfuse.com/docs/experimentation), [Prompt Management](https://langfuse.com/docs/prompts))
+- [Evaluations](https://langfuse.com/docs/scores/overview) are key to the LLM application development workflow, and Langfuse adapts to your needs. It supports LLM-as-a-judge, user feedback collection, manual labeling, and custom evaluation pipelines via APIs/SDKs.
 
-## Get started
+- [Datasets](https://langfuse.com/docs/datasets/overview) enable test sets and benchmarks for evaluating your LLM application. They support continuous improvement, pre-deployment testing, structured experiments, flexible evaluation, and seamless integration with frameworks like LangChain and LlamaIndex.
+
+- [LLM Playground](https://langfuse.com/docs/playground) is a tool for testing and iterating on your prompts and model configurations, shortening the feedback loop and accelerating development. When you see a bad result in tracing, you can directly jump to the playground to iterate on it.
+
+- [Comprehensive API](https://langfuse.com/docs/api): Langfuse is frequently used to power bespoke LLMOps workflows while using the building blocks provided by Langfuse via the API. OpenAPI spec, Postman collection, and typed SDKs for Python, JS/TS are available.
+
+## 📦 Deploy Langfuse
+
+![Langfuse Deployment Options](https://langfuse.com/images/docs/github-readme/github-deployment-options.png)
 
 ### Langfuse Cloud
 
 Managed deployment by the Langfuse team, generous free-tier (hobby plan), no credit card required.
 
-**[» Langfuse Cloud](https://cloud.langfuse.com)**
+<div align="center">
+    <a href="https://cloud.langfuse.com" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/»%20Sign%20up%20for%20Langfuse%20Cloud-8A2BE2?&color=orange">
+    </a>
+</div>
 
-## Self-Hosting Open Source LLM Observability with Langfuse
+### Self-Host Langfuse
 
-### Localhost (docker)
+Run Langfuse on your own infrastructure:
 
-```bash
-# Clone repository
-git clone https://github.com/langfuse/langfuse.git
-cd langfuse
+- [Local (docker compose)](https://langfuse.com/self-hosting/local): Run Langfuse on your own machine in 5 minutes using Docker Compose.
 
-# Run server and database
-docker compose up -d
-```
+  ```bash
+  # Get a copy of the latest Langfuse repository
+  git clone https://github.com/langfuse/langfuse.git
+  cd langfuse
 
-[→ Learn more about deploying locally](https://langfuse.com/docs/deployment/local)
+  # Run the langfuse docker compose
+  docker compose up
+  ```
 
-### Self-host (docker)
+- [Kubernetes (Helm)](https://langfuse.com/self-hosting/kubernetes-helm): Run Langfuse on a Kubernetes cluster using Helm. This is the preferred production deployment.
+- [VM](https://langfuse.com/self-hosting/docker-compose): Run Langfuse on a single Virtual Machine using Docker Compose.
+- Planned: Cloud-specific deployment guides, please upvote and comment on the following threads: [AWS](https://github.com/orgs/langfuse/discussions/4645), [Google Cloud](https://github.com/orgs/langfuse/discussions/4646), [Azure](https://github.com/orgs/langfuse/discussions/4647).
 
-Langfuse is simple to self-host and keep updated. It currently requires only a single docker container and a postgres database.
-[→ Self Hosting Instructions](https://langfuse.com/docs/deployment/self-host)
+See [self-hosting documentation](https://langfuse.com/self-hosting) to learn more about the architecture and configuration options.
 
-Templated deployments: [Railway, GCP, AWS, Azure, Kubernetes and others](https://langfuse.com/docs/deployment/self-host#platform-specific-information)
+## 🔌 Integrations
 
-## Get Started
+![Langfuse Integrations](https://langfuse.com/images/docs/github-readme/github-integrations.png)
 
-### API Keys
-
-You need a Langfuse public and secret key to get started. Sign up [here](https://cloud.langfuse.com) and find them in your project settings.
-
-### Ingesting Data · Instrumenting Your Application · LLM Observability with Langfuse
-
-Note: We recommend using our fully async, typed [SDKs](https://langfuse.com/docs/sdk) that allow you to instrument any LLM application with any underlying model. They are available in [Python (Decorators)](https://langfuse.com/docs/sdk/python) & [JS/TS](https://langfuse.com/docs/sdk/typescript). The SDKs will always be the most fully featured and stable way to ingest data into Langfuse.
-
-See the [→ Quickstart](https://langfuse.com/docs/get-started) to integrate Langfuse.
-
-### LLM Observability Integrations
+### Main Integrations:
 
 | Integration                                                                  | Supports                   | Description                                                                                                                                      |
 | ---------------------------------------------------------------------------- | -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ |
@@ -122,50 +139,180 @@ See the [→ Quickstart](https://langfuse.com/docs/get-started) to integrate Lan
 | [Vercel AI SDK](https://langfuse.com/docs/integrations/vercel-ai-sdk)        | JS/TS                      | TypeScript toolkit designed to help developers build AI-powered applications with React, Next.js, Vue, Svelte, Node.js.                          |
 | [API](https://langfuse.com/docs/api)                                         |                            | Directly call the public API. OpenAPI spec available.                                                                                            |
 
-Packages integrated with Langfuse:
+### Packages integrated with Langfuse:
 
-| Name                                                            | Description                                                                                                             |
-| --------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
-| [Instructor](https://langfuse.com/docs/integrations/instructor) | Library to get structured LLM outputs (JSON, Pydantic)                                                                  |
-| [Dify](https://langfuse.com/docs/integrations/dify)             | Open source LLM app development platform with no-code builder.                                                          |
-| [Ollama](https://langfuse.com/docs/integrations/ollama)         | Easily run open source LLMs on your own machine.                                                                        |
-| [Mirascope](https://langfuse.com/docs/integrations/mirascope)   | Python toolkit for building LLM applications.                                                                           |
-| [Flowise](https://langfuse.com/docs/integrations/flowise)       | JS/TS no-code builder for customized LLM flows.                                                                         |
-| [Langflow](https://langfuse.com/docs/integrations/langflow)     | Python-based UI for LangChain, designed with react-flow to provide an effortless way to experiment and prototype flows. |
+| Name                                                                    | Type               | Description                                                                                                             |
+| ----------------------------------------------------------------------- | ------------------ | ----------------------------------------------------------------------------------------------------------------------- |
+| [Instructor](https://langfuse.com/docs/integrations/instructor)         | Library            | Library to get structured LLM outputs (JSON, Pydantic)                                                                  |
+| [DSPy](https://langfuse.com/docs/integrations/dspy)                     | Library            | Framework that systematically optimizes language model prompts and weights                                              |
+| [Mirascope](https://langfuse.com/docs/integrations/mirascope)           | Library            | Python toolkit for building LLM applications.                                                                           |
+| [Ollama](https://langfuse.com/docs/integrations/ollama)                 | Model (local)      | Easily run open source LLMs on your own machine.                                                                        |
+| [Amazon Bedrock](https://langfuse.com/docs/integrations/amazon-bedrock) | Model              | Run foundation and fine-tuned models on AWS.                                                                            |
+| [AutoGen](https://langfuse.com/docs/integrations/autogen)               | Agent Framework    | Open source LLM platform for building distributed agents.                                                               |
+| [Flowise](https://langfuse.com/docs/integrations/flowise)               | Chat/Agent&nbsp;UI | JS/TS no-code builder for customized LLM flows.                                                                         |
+| [Langflow](https://langfuse.com/docs/integrations/langflow)             | Chat/Agent&nbsp;UI | Python-based UI for LangChain, designed with react-flow to provide an effortless way to experiment and prototype flows. |
+| [Dify](https://langfuse.com/docs/integrations/dify)                     | Chat/Agent&nbsp;UI | Open source LLM app development platform with no-code builder.                                                          |
+| [OpenWebUI](https://langfuse.com/docs/integrations/openwebui)           | Chat/Agent&nbsp;UI | Self-hosted LLM Chat web ui supporting various LLM runners including self-hosted and local models.                      |
+| [Promptfoo](https://langfuse.com/docs/integrations/promptfoo)           | Tool               | Open source LLM testing platform.                                                                                       |
+| [LobeChat](https://langfuse.com/docs/integrations/lobechat)             | Chat/Agent&nbsp;UI | Open source chatbot platform.                                                                                           |
+| [Vapi](https://langfuse.com/docs/integrations/vapi)                     | Platform           | Open source voice AI platform.                                                                                          |
+| [Inferable](https://langfuse.com/docs/integrations/other/inferable)     | Agents             | Open source LLM platform for building distributed agents.                                                               |
+| [Gradio](https://langfuse.com/docs/integrations/other/gradio)           | Chat/Agent&nbsp;UI | Open source Python library to build web interfaces like Chat UI.                                                        |
+| [Goose](https://langfuse.com/docs/integrations/goose)                   | Agents             | Open source LLM platform for building distributed agents.                                                               |
+| [smolagents](https://langfuse.com/docs/integrations/smolagents)         | Agents             | Open source AI agents framework.                                                                                        |
+| [CrewAI](https://langfuse.com/docs/integrations/crewai)                 | Agents             | Multi agent framework for agent collaboration and tool use.                                                             |
 
-## Questions and feedback
+## 🚀 Quickstart
 
-### Ideas and roadmap
+Instrument your app and start ingesting traces to Langfuse, thereby tracking LLM calls and other relevant logic in your app such as retrieval, embedding, or agent actions. Inspect and debug complex logs and user sessions.
 
-- [Roadmap](https://langfuse.com/roadmap)
-- [GitHub Discussions](https://github.com/orgs/langfuse/discussions)
-- [Feature Requests](https://langfuse.com/ideas)
+### 1️⃣ Create new project
 
-### Support and feedback
+1.  [Create Langfuse account](https://cloud.langfuse.com/auth/sign-up) or [self-host](https://langfuse.com/self-hosting)
+2.  Create a new project
+3.  Create new API credentials in the project settings
 
-In order of preference the best way to communicate with us:
+### 2️⃣ Log your first LLM call
 
-- [GitHub Discussions](https://github.com/orgs/langfuse/discussions) (preferred): Contribute [ideas](https://langfuse.com/ideas), [support requests](https://langfuse.com/gh-support) and [report bugs](https://langfuse.com/issues)
-- [Discord](https://langfuse.com/discord): community support
-- Privately: contact at langfuse dot com
+The [`@observe()` decorator](https://langfuse.com/docs/sdk/python/decorators) makes it easy to trace any Python LLM application. In this quickstart we also use the Langfuse [OpenAI integration](https://langfuse.com/docs/integrations/openai) to automatically capture all model parameters.
 
-## Contributing to Langfuse
+> [!TIP]
+> Not using OpenAI? Visit [our documentation](https://langfuse.com/docs/get-started#log-your-first-llm-call-to-langfuse) to learn how to log other models and frameworks.
 
-- Vote on [Ideas](https://github.com/orgs/langfuse/discussions/categories/ideas)
-- Raise and comment on [Issues](https://github.com/langfuse/langfuse/issues)
+```bash
+pip install langfuse openai
+```
+
+```bash filename=".env"
+LANGFUSE_SECRET_KEY="sk-lf-..."
+LANGFUSE_PUBLIC_KEY="pk-lf-..."
+LANGFUSE_HOST="https://cloud.langfuse.com" # 🇪🇺 EU region
+# LANGFUSE_HOST="https://us.cloud.langfuse.com" # 🇺🇸 US region
+```
+
+```python /@observe()/ /from langfuse.openai import openai/ filename="main.py"
+from langfuse.decorators import observe
+from langfuse.openai import openai # OpenAI integration
+
+@observe()
+def story():
+    return openai.chat.completions.create(
+        model="gpt-4o",
+        messages=[{"role": "user", "content": "What is Langfuse?"}],
+    ).choices[0].message.content
+
+@observe()
+def main():
+    return story()
+
+main()
+```
+
+### 3️⃣ See traces in Langfuse
+
+See your language model calls and other application logic in Langfuse.
+
+![Example trace in Langfuse](https://langfuse.com/images/docs/github-readme/github-example-trace.png)
+
+_[Public example trace in Langfuse](https://cloud.langfuse.com/project/cloramnkj0002jz088vzn1ja4/traces/2cec01e3-3dc2-472f-afcf-3b968cf0c1f4?timestamp=2025-02-10T14%3A27%3A30.275Z&observation=cb5ff844-07ef-41e6-b8e2-6c64344bc13b)_
+
+> [!TIP]
+>
+> [Learn more](https://langfuse.com/docs/tracing) about tracing in Langfuse or play with the [interactive demo](https://langfuse.com/docs/demo).
+
+## ⭐️ Star Us
+
+![Star Langfuse](https://langfuse.com/images/docs/github-readme/github-star-howto.gif)
+
+## 💭 Support
+
+Finding an answer to your question:
+
+- Our [documentation](https://langfuse.com/docs) is the best place to start looking for answers. It is comprehensive, and we invest significant time into maintaining it. You can also suggest edits to the docs via GitHub.
+- [Langfuse FAQs](https://langfuse.com/faq) where the most common questions are answered.
+- Use "[Ask AI](https://langfuse.com/docs/ask-ai)" to get instant answers to your questions.
+
+Support Channels:
+
+- **Ask any question in our [public Q&A](https://github.com/orgs/langfuse/discussions/categories/support) on GitHub Discussions.** Please include as much detail as possible (e.g. code snippets, screenshots, background information) to help us understand your question.
+- [Request a feature](https://github.com/orgs/langfuse/discussions/categories/ideas) on GitHub Discussions.
+- [Report a Bug](https://github.com/langfuse/langfuse/issues) on GitHub Issues.
+- For time-sensitive queries, ping us via the in-app chat widget.
+
+## 🤝 Contributing
+
+Your contributions are welcome!
+
+- Vote on [Ideas](https://github.com/orgs/langfuse/discussions/categories/ideas) in GitHub Discussions.
+- Raise and comment on [Issues](https://github.com/langfuse/langfuse/issues).
 - Open a PR - see [CONTRIBUTING.md](CONTRIBUTING.md) for details on how to setup a development environment.
 
-## License
+## 🥇 License
 
 This repository is MIT licensed, except for the `ee` folders. See [LICENSE](LICENSE) and [docs](https://langfuse.com/docs/open-source) for more details.
 
-## Misc
+## ⭐️ Star History
 
-### GET API to export your data
+<a href="https://star-history.com/#langfuse/langfuse&Date">
+ <picture>
+   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date&theme=dark" />
+   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
+   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" style="border-radius: 15px;" />
+ </picture>
+</a>
 
-[**GET routes**](https://langfuse.com/docs/integrations/api) to use data in downstream applications (e.g. embedded analytics). You can also access them conveniently via the SDKs ([docs](https://langfuse.com/docs/query-traces)).
+## ❤️ Open Source Projects Using Langfuse
 
-### Security & Privacy
+Top open-source Python projects that use Langfuse, ranked by stars ([Source](https://github.com/langfuse/langfuse-docs/blob/main/components-mdx/dependents)):
+
+| Repository                                                                                                                                                                                                                                                          | Stars |
+| :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----: |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/127165244?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langgenius](https://github.com/langgenius) / [dify](https://github.com/langgenius/dify)                                            | 54865 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [open-webui](https://github.com/open-webui/open-webui)                                | 51531 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/131470832?s=40&v=4" width="20" height="20" alt=""> &nbsp; [lobehub](https://github.com/lobehub) / [lobe-chat](https://github.com/lobehub/lobe-chat)                                           | 49003 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85702467?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langflow-ai](https://github.com/langflow-ai) / [langflow](https://github.com/langflow-ai/langflow)                                  | 39093 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/130722866?s=40&v=4" width="20" height="20" alt=""> &nbsp; [run-llama](https://github.com/run-llama) / [llama_index](https://github.com/run-llama/llama_index)                                 | 37368 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/139558948?s=40&v=4" width="20" height="20" alt=""> &nbsp; [chatchat-space](https://github.com/chatchat-space) / [Langchain-Chatchat](https://github.com/chatchat-space/Langchain-Chatchat)    | 32486 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/128289781?s=40&v=4" width="20" height="20" alt=""> &nbsp; [FlowiseAI](https://github.com/FlowiseAI) / [Flowise](https://github.com/FlowiseAI/Flowise)                                         | 32448 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/31035808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mindsdb](https://github.com/mindsdb) / [mindsdb](https://github.com/mindsdb/mindsdb)                                                | 26931 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/119600397?s=40&v=4" width="20" height="20" alt=""> &nbsp; [twentyhq](https://github.com/twentyhq) / [twenty](https://github.com/twentyhq/twenty)                                              | 24195 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog](https://github.com/PostHog/posthog)                                                | 22618 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/121462774?s=40&v=4" width="20" height="20" alt=""> &nbsp; [BerriAI](https://github.com/BerriAI) / [litellm](https://github.com/BerriAI/litellm)                                               | 15151 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/179202840?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mediar-ai](https://github.com/mediar-ai) / [screenpipe](https://github.com/mediar-ai/screenpipe)                                   | 11037 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/105877416?s=40&v=4" width="20" height="20" alt=""> &nbsp; [formbricks](https://github.com/formbricks) / [formbricks](https://github.com/formbricks/formbricks)                                |  9386 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/76263028?s=40&v=4" width="20" height="20" alt=""> &nbsp; [anthropics](https://github.com/anthropics) / [courses](https://github.com/anthropics/courses)                                       |  8385 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/78410652?s=40&v=4" width="20" height="20" alt=""> &nbsp; [GreyDGL](https://github.com/GreyDGL) / [PentestGPT](https://github.com/GreyDGL/PentestGPT)                                          |  7374 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/152537519?s=40&v=4" width="20" height="20" alt=""> &nbsp; [superagent-ai](https://github.com/superagent-ai) / [superagent](https://github.com/superagent-ai/superagent)                       |  5391 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/137907881?s=40&v=4" width="20" height="20" alt=""> &nbsp; [promptfoo](https://github.com/promptfoo) / [promptfoo](https://github.com/promptfoo/promptfoo)                                     |  4976 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/157326433?s=40&v=4" width="20" height="20" alt=""> &nbsp; [onlook-dev](https://github.com/onlook-dev) / [onlook](https://github.com/onlook-dev/onlook)                                        |  4141 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/7250217?s=40&v=4" width="20" height="20" alt=""> &nbsp; [Canner](https://github.com/Canner) / [WrenAI](https://github.com/Canner/WrenAI)                                                      |  2526 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/11855343?s=40&v=4" width="20" height="20" alt=""> &nbsp; [pingcap](https://github.com/pingcap) / [autoflow](https://github.com/pingcap/autoflow)                                              |  2061 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85268109?s=40&v=4" width="20" height="20" alt=""> &nbsp; [MLSysOps](https://github.com/MLSysOps) / [MLE-agent](https://github.com/MLSysOps/MLE-agent)                                         |  1161 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [pipelines](https://github.com/open-webui/pipelines)                                  |  1100 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/18422723?s=40&v=4" width="20" height="20" alt=""> &nbsp; [alishobeiri](https://github.com/alishobeiri) / [thread](https://github.com/alishobeiri/thread)                                      |  1074 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/125468716?s=40&v=4" width="20" height="20" alt=""> &nbsp; [topoteretes](https://github.com/topoteretes) / [cognee](https://github.com/topoteretes/cognee)                                     |   971 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/188657705?s=40&v=4" width="20" height="20" alt=""> &nbsp; [bRAGAI](https://github.com/bRAGAI) / [bRAG-langchain](https://github.com/bRAGAI/bRAG-langchain)                                    |   823 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169500408?s=40&v=4" width="20" height="20" alt=""> &nbsp; [opslane](https://github.com/opslane) / [opslane](https://github.com/opslane/opslane)                                               |   677 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/151867818?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dynamiq-ai](https://github.com/dynamiq-ai) / [dynamiq](https://github.com/dynamiq-ai/dynamiq)                                      |   639 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/48585267?s=40&v=4" width="20" height="20" alt=""> &nbsp; [theopenconversationkit](https://github.com/theopenconversationkit) / [tock](https://github.com/theopenconversationkit/tock)         |   514 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/20493493?s=40&v=4" width="20" height="20" alt=""> &nbsp; [andysingal](https://github.com/andysingal) / [llm-course](https://github.com/andysingal/llm-course)                                 |   394 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/132396805?s=40&v=4" width="20" height="20" alt=""> &nbsp; [phospho-app](https://github.com/phospho-app) / [phospho](https://github.com/phospho-app/phospho)                                   |   384 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/178644984?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sentient-engineering](https://github.com/sentient-engineering) / [agent-q](https://github.com/sentient-engineering/agent-q)        |   370 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/168552753?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sql-agi](https://github.com/sql-agi) / [DB-GPT](https://github.com/sql-agi/DB-GPT)                                                 |   324 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog-foss](https://github.com/PostHog/posthog-foss)                                      |   305 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/154247157?s=40&v=4" width="20" height="20" alt=""> &nbsp; [vespperhq](https://github.com/vespperhq) / [vespper](https://github.com/vespperhq/vespper)                                         |   304 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/185116535?s=40&v=4" width="20" height="20" alt=""> &nbsp; [block](https://github.com/block) / [goose](https://github.com/block/goose)                                                         |   295 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/609489?s=40&v=4" width="20" height="20" alt=""> &nbsp; [aorwall](https://github.com/aorwall) / [moatless-tools](https://github.com/aorwall/moatless-tools)                                    |   291 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/2357342?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dmayboroda](https://github.com/dmayboroda) / [minima](https://github.com/dmayboroda/minima)                                          |   221 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/66303003?s=40&v=4" width="20" height="20" alt=""> &nbsp; [RobotecAI](https://github.com/RobotecAI) / [rai](https://github.com/RobotecAI/rai)                                                  |   172 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/148684274?s=40&v=4" width="20" height="20" alt=""> &nbsp; [i-am-alice](https://github.com/i-am-alice) / [3rd-devs](https://github.com/i-am-alice/3rd-devs)                                    |   148 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/171735272?s=40&v=4" width="20" height="20" alt=""> &nbsp; [8090-inc](https://github.com/8090-inc) / [xrx-sample-apps](https://github.com/8090-inc/xrx-sample-apps)                            |   138 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/104478511?s=40&v=4" width="20" height="20" alt=""> &nbsp; [babelcloud](https://github.com/babelcloud) / [LLM-RGB](https://github.com/babelcloud/LLM-RGB)                                      |   135 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/15125613?s=40&v=4" width="20" height="20" alt=""> &nbsp; [souzatharsis](https://github.com/souzatharsis) / [tamingLLMs](https://github.com/souzatharsis/tamingLLMs)                           |   129 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169401942?s=40&v=4" width="20" height="20" alt=""> &nbsp; [LibreChat-AI](https://github.com/LibreChat-AI) / [librechat.ai](https://github.com/LibreChat-AI/librechat.ai)                      |   128 |
+| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/51827949?s=40&v=4" width="20" height="20" alt=""> &nbsp; [deepset-ai](https://github.com/deepset-ai) / [haystack-core-integrations](https://github.com/deepset-ai/haystack-core-integrations) |   126 |
+
+## 🔒 Security & Privacy
 
 We take data security and privacy seriously. Please refer to our [Security and Privacy](https://langfuse.com/security) page for more information.
 
@@ -181,50 +328,3 @@ This helps us to:
 None of the data is shared with third parties and does not include any sensitive information. We want to be super transparent about this and you can find the exact data we collect [here](/web/src/features/telemetry/index.ts).
 
 You can opt-out by setting `TELEMETRY_ENABLED=false`.
-
-### Star History
-
-<a href="https://star-history.com/#langfuse/langfuse&Date">
- <picture>
-   <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date&theme=dark" />
-   <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
-   <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=langfuse/langfuse&type=Date" />
- </picture>
-</a>
-
-### Open Source Projects Using Langfuse
-
-Top open-source Python projects that use Langfuse, ranked by stars ([Source](https://github.com/langfuse/langfuse-docs/blob/main/components-mdx/dependents/python.md)):
-
-| Repository                                                                                                                                                                                                                                                          | Stars |
-| :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----: |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/127165244?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langgenius](https://github.com/langgenius) / [dify](https://github.com/langgenius/dify)                                            | 53464 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [open-webui](https://github.com/open-webui/open-webui)                                | 49387 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/130722866?s=40&v=4" width="20" height="20" alt=""> &nbsp; [run-llama](https://github.com/run-llama) / [llama_index](https://github.com/run-llama/llama_index)                                 | 37087 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85702467?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langflow-ai](https://github.com/langflow-ai) / [langflow](https://github.com/langflow-ai/langflow)                                  | 36690 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/31035808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [mindsdb](https://github.com/mindsdb) / [mindsdb](https://github.com/mindsdb/mindsdb)                                                | 26871 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog](https://github.com/PostHog/posthog)                                                | 22348 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/121462774?s=40&v=4" width="20" height="20" alt=""> &nbsp; [BerriAI](https://github.com/BerriAI) / [litellm](https://github.com/BerriAI/litellm)                                               | 14508 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/78410652?s=40&v=4" width="20" height="20" alt=""> &nbsp; [GreyDGL](https://github.com/GreyDGL) / [PentestGPT](https://github.com/GreyDGL/PentestGPT)                                          |  7319 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/152537519?s=40&v=4" width="20" height="20" alt=""> &nbsp; [superagent-ai](https://github.com/superagent-ai) / [superagent](https://github.com/superagent-ai/superagent)                       |  5352 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/7250217?s=40&v=4" width="20" height="20" alt=""> &nbsp; [Canner](https://github.com/Canner) / [WrenAI](https://github.com/Canner/WrenAI)                                                      |  2084 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/11855343?s=40&v=4" width="20" height="20" alt=""> &nbsp; [pingcap](https://github.com/pingcap) / [autoflow](https://github.com/pingcap/autoflow)                                              |  1936 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/85268109?s=40&v=4" width="20" height="20" alt=""> &nbsp; [MLSysOps](https://github.com/MLSysOps) / [MLE-agent](https://github.com/MLSysOps/MLE-agent)                                         |  1140 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/158137808?s=40&v=4" width="20" height="20" alt=""> &nbsp; [open-webui](https://github.com/open-webui) / [pipelines](https://github.com/open-webui/pipelines)                                  |  1044 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/125468716?s=40&v=4" width="20" height="20" alt=""> &nbsp; [topoteretes](https://github.com/topoteretes) / [cognee](https://github.com/topoteretes/cognee)                                     |   945 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/188657705?s=40&v=4" width="20" height="20" alt=""> &nbsp; [bRAGAI](https://github.com/bRAGAI) / [bRAG-langchain](https://github.com/bRAGAI/bRAG-langchain)                                    |   770 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/169500408?s=40&v=4" width="20" height="20" alt=""> &nbsp; [opslane](https://github.com/opslane) / [opslane](https://github.com/opslane/opslane)                                               |   675 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/151867818?s=40&v=4" width="20" height="20" alt=""> &nbsp; [dynamiq-ai](https://github.com/dynamiq-ai) / [dynamiq](https://github.com/dynamiq-ai/dynamiq)                                      |   583 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/48585267?s=40&v=4" width="20" height="20" alt=""> &nbsp; [theopenconversationkit](https://github.com/theopenconversationkit) / [tock](https://github.com/theopenconversationkit/tock)         |   505 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/20493493?s=40&v=4" width="20" height="20" alt=""> &nbsp; [andysingal](https://github.com/andysingal) / [llm-course](https://github.com/andysingal/llm-course)                                 |   383 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/132396805?s=40&v=4" width="20" height="20" alt=""> &nbsp; [phospho-app](https://github.com/phospho-app) / [phospho](https://github.com/phospho-app/phospho)                                   |   379 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/178644984?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sentient-engineering](https://github.com/sentient-engineering) / [agent-q](https://github.com/sentient-engineering/agent-q)        |   357 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/168552753?s=40&v=4" width="20" height="20" alt=""> &nbsp; [sql-agi](https://github.com/sql-agi) / [DB-GPT](https://github.com/sql-agi/DB-GPT)                                                 |   324 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/60330232?s=40&v=4" width="20" height="20" alt=""> &nbsp; [PostHog](https://github.com/PostHog) / [posthog-foss](https://github.com/PostHog/posthog-foss)                                      |   300 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/609489?s=40&v=4" width="20" height="20" alt=""> &nbsp; [aorwall](https://github.com/aorwall) / [moatless-tools](https://github.com/aorwall/moatless-tools)                                    |   278 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/185116535?s=40&v=4" width="20" height="20" alt=""> &nbsp; [block](https://github.com/block) / [goose](https://github.com/block/goose)                                                         |   237 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/66303003?s=40&v=4" width="20" height="20" alt=""> &nbsp; [RobotecAI](https://github.com/RobotecAI) / [rai](https://github.com/RobotecAI/rai)                                                  |   164 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/171735272?s=40&v=4" width="20" height="20" alt=""> &nbsp; [8090-inc](https://github.com/8090-inc) / [xrx-sample-apps](https://github.com/8090-inc/xrx-sample-apps)                            |   133 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/51827949?s=40&v=4" width="20" height="20" alt=""> &nbsp; [deepset-ai](https://github.com/deepset-ai) / [haystack-core-integrations](https://github.com/deepset-ai/haystack-core-integrations) |   122 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/134601687?s=40&v=4" width="20" height="20" alt=""> &nbsp; [langfuse](https://github.com/langfuse) / [langfuse-python](https://github.com/langfuse/langfuse-python)                            |   108 |
-| <img class="avatar mr-2" src="https://avatars.githubusercontent.com/u/42655961?s=40&v=4" width="20" height="20" alt=""> &nbsp; [zozoheir](https://github.com/zozoheir) / [tinyllm](https://github.com/zozoheir/tinyllm)                                             |    95 |

careers

@@ -1,13 +0,0 @@
-_We are Hiring_
-Join us in scaling Langfuse in Berlin, Germany. We are an open source company, we hire in person, we are only hiring technical talent.
-
-_Open Roles_
-
-- Product Engineer, 70-130k EUR, 0.25-0.75% Equity, https://www.ycombinator.com/companies/langfuse/jobs/aAvmoFB-product-engineer
-- Developer Advocate, 60-110k EUR, 0.25-0.5% Equity, https://www.ycombinator.com/companies/langfuse/jobs/uHysbKH-developer-advocate-devrel
-
-_More Info_
-
-- https://langfuse.com/careers
-- https://langfuse.com/docs
-- https://langfuse.com/changelog

docker-compose.build.yml

@@ -155,4 +155,3 @@ volumes:
     driver: local
   langfuse_minio_data:
     driver: local
-

docker-compose.yml

@@ -38,6 +38,8 @@ services:
       LANGFUSE_S3_MEDIA_UPLOAD_ENDPOINT: ${LANGFUSE_S3_MEDIA_UPLOAD_ENDPOINT:-http://minio:9000}
       LANGFUSE_S3_MEDIA_UPLOAD_FORCE_PATH_STYLE: ${LANGFUSE_S3_MEDIA_UPLOAD_FORCE_PATH_STYLE:-true}
       LANGFUSE_S3_MEDIA_UPLOAD_PREFIX: ${LANGFUSE_S3_MEDIA_UPLOAD_PREFIX:-media/}
+      LANGFUSE_INGESTION_QUEUE_DELAY_MS: ${LANGFUSE_INGESTION_QUEUE_DELAY_MS:-}
+      LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS: ${LANGFUSE_INGESTION_CLICKHOUSE_WRITE_INTERVAL_MS:-}
       REDIS_HOST: ${REDIS_HOST:-redis}
       REDIS_PORT: ${REDIS_PORT:-6379}
       REDIS_AUTH: ${REDIS_AUTH:-myredissecret}
@@ -61,10 +63,6 @@ services:
       LANGFUSE_INIT_USER_EMAIL: ${LANGFUSE_INIT_USER_EMAIL:-}
       LANGFUSE_INIT_USER_NAME: ${LANGFUSE_INIT_USER_NAME:-}
       LANGFUSE_INIT_USER_PASSWORD: ${LANGFUSE_INIT_USER_PASSWORD:-}
-      LANGFUSE_SDK_CI_SYNC_PROCESSING_ENABLED: ${LANGFUSE_SDK_CI_SYNC_PROCESSING_ENABLED:-false}
-      LANGFUSE_READ_FROM_POSTGRES_ONLY: ${LANGFUSE_READ_FROM_POSTGRES_ONLY:-false}
-      LANGFUSE_READ_FROM_CLICKHOUSE_ONLY: ${LANGFUSE_READ_FROM_CLICKHOUSE_ONLY:-true}
-      LANGFUSE_RETURN_FROM_CLICKHOUSE: ${LANGFUSE_RETURN_FROM_CLICKHOUSE:-true}
 
   clickhouse:
     image: clickhouse/clickhouse-server

ee/LICENSE

@@ -1,24 +1,36 @@
-Langfuse Enterprise License (the “Enterprise License”)
-Copyright (c) 2024 Langfuse GmbH ('Langfuse')
+START NOTE
+Langfuse is an open core project. Langfuse's core
+is permissively licensed (MIT license).
+Certain parts of the periphery of Langfuse are commercially
+licensed and governed by this Enterprise License.
+For the avoidance of doubt, this license does not apply 
+to the core of Langfuse as it is defined in its license in 
+the directory "/LICENSE" (as opposed to this file  "ee/LICENSE")
+which can be used and run without infringing the below license
+and its licensed materials.
+END NOTE
 
-With regard to the Langfuse Software:
+START OF LICENSE
+Langfuse Enterprise license (the “Enterprise License” or "EE license")
+
+Copyright (c) 2023-2025 Langfuse GmbH
+
+With regard to the Langfuse Enterprise Software:
 This software and associated documentation files (the "Software") may only be
-used in production, if you (and any entity that you represent) have agreed to,
-and are in compliance with, the Langfuse Terms of Service, available
+used, if you (and any entity that you represent) have agreed to,
+and are in compliance with, the applicable Langfuse Terms of Service, available
 at https://langfuse.com/terms (the “Enterprise Terms”), or other
 agreement governing the use of the Software, as agreed by you and Langfuse,
-and otherwise have a valid Langfuse Enterprise license.
+and otherwise have a valid Langfuse Enterprise License.
 
 Subject to the foregoing sentence, you are free to
 modify this Software and publish patches to the Software. You agree that Langfuse
 and/or its licensors (as applicable) retain all right, title and interest in and
 to all such modifications and/or patches, and all such modifications and/or
 patches may only be used, copied, modified, displayed, distributed, or otherwise
-exploited with a valid Langfuse Enterprise license.
-
-Notwithstanding the foregoing, you may copy and modify
+exploited with a valid Langfuse Enterprise License. Notwithstanding the foregoing, you may copy and modify
 the Software for development and testing purposes, without requiring a
-subscription. You agree that Langfuse and/or its licensors (as applicable) retain
+subscription. You agree that Langfuse GmbH and/or its licensors (as applicable) retain
 all right, title and interest in and to all such modifications. You are not
 granted any other rights beyond what is expressly stated herein. Subject to the
 foregoing, it is forbidden to copy, merge, publish, distribute, sublicense,
@@ -33,5 +45,7 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 
 For all third party components incorporated into the Langfuse Software, those
-components are licensed under the original license provided by the owner of the 
+components are licensed under the original license provided by the owner of the
 applicable component.
+
+END OF LICENSE

ee/package.json

@@ -27,7 +27,8 @@
     "@langfuse/shared": "workspace:*",
     "@opentelemetry/api": ">=1.0.0 <1.10.0",
     "axios": "^1.7.7",
-    "next": "^14.2.15",
+    "https-proxy-agent": "^7.0.6",
+    "next": "^14.2.21",
     "next-auth": "^4.24.11",
     "zod": "^3.23.8"
   },
@@ -47,7 +48,7 @@
   },
   "pnpm": {
     "overrides": {
-      "jsonpath-plus": "10.0.7"
+      "nanoid": "^3.3.8"
     }
   }
 }

fern/apis/client/generators.yml

@@ -12,7 +12,7 @@ groups:
       #     namespaceExport: Langfuse
       #     allowCustomFetcher: true
       - name: fernapi/fern-openapi
-        version: 0.0.26
+        version: 0.1.7
         output:
           location: local-file-system
           path: ../../../web/public/generated/api-client

fern/apis/server/definition/commons.yml

@@ -378,10 +378,10 @@ types:
         type: string
       startDate:
         docs: Apply only to generations which are newer than this ISO date.
-        type: optional<date>
+        type: optional<datetime>
       unit:
         docs: Unit used by this model.
-        type: ModelUsageUnit
+        type: optional<ModelUsageUnit>
       inputPrice:
         docs: Price (USD) per input unit
         type: optional<double>

fern/apis/server/definition/ingestion.yml

@@ -11,6 +11,7 @@ service:
         Batched ingestion for Langfuse Tracing. If you want to use tracing via the API, such as to build your own Langfuse client implementation, this is the only API route you need to implement.
 
         Notes:
+        - Introduction to data model: https://langfuse.com/docs/tracing-data-model
         - Batch sizes are limited to 3.5 MB in total. You need to adjust the number of events per batch accordingly.
         - The API does not return a 4xx status code for input errors. Instead, it responds with a 207 status code, which includes a list of the encountered errors.
       method: POST
@@ -127,7 +128,7 @@ types:
       modelParameters: optional<map<string, commons.MapValue>>
       usage: optional<IngestionUsage>
       usageDetails: optional<UsageDetails>
-      costDetails: optional<map<string, float>>
+      costDetails: optional<map<string, double>>
       promptName: optional<string>
       promptVersion: optional<integer>
 
@@ -140,7 +141,7 @@ types:
       usage: optional<IngestionUsage>
       promptName: optional<string>
       usageDetails: optional<UsageDetails>
-      costDetails: optional<map<string, float>>
+      costDetails: optional<map<string, double>>
       promptVersion: optional<integer>
 
   ObservationBody:

fern/apis/server/definition/models.yml

@@ -58,7 +58,7 @@ types:
         type: optional<datetime>
       unit:
         docs: Unit used by this model.
-        type: commons.ModelUsageUnit
+        type: optional<commons.ModelUsageUnit>
       inputPrice:
         docs: Price (USD) per input unit
         type: optional<double>

fern/apis/server/definition/prompt-version.yml

@@ -0,0 +1,27 @@
+# yaml-language-server: $schema=https://raw.githubusercontent.com/fern-api/fern/main/fern.schema.json
+imports:
+  prompts: ./prompts.yml
+  pagination: ./utils/pagination.yml
+service:
+  auth: true
+  base-path: /api/public/v2
+  endpoints:
+    update:
+      docs: Update labels for a specific prompt version
+      method: PATCH
+      path: /prompts/{name}/versions/{version}
+      path-parameters:
+        name:
+          type: string
+          docs: The name of the prompt
+        version:
+          type: integer
+          docs: Version of the prompt to update
+      request:
+        name: UpdatePromptRequest
+        body:
+          properties:
+            newLabels:
+              type: list<string>
+              docs: New labels for the prompt version. Labels are unique across versions. The "latest" label is reserved and managed by Langfuse.
+      response: prompts.Prompt

fern/apis/server/definition/prompts.yml

@@ -90,6 +90,9 @@ types:
       tags:
         type: optional<list<string>>
         docs: List of tags to apply to all versions of this prompt.
+      commitMessage:
+        type: optional<string>
+        docs: Commit message for this prompt version.
 
   CreateTextPromptRequest:
     properties:
@@ -102,6 +105,9 @@ types:
       tags:
         type: optional<list<string>>
         docs: List of tags to apply to all versions of this prompt.
+      commitMessage:
+        type: optional<string>
+        docs: Commit message for this prompt version.
 
   Prompt:
     union:
@@ -119,6 +125,9 @@ types:
       tags:
         type: list<string>
         docs: List of tags. Used to filter via UI and API. The same across versions of a prompt.
+      commitMessage:
+        type: optional<string>
+        docs: Commit message for this prompt version.
 
   ChatMessage:
     properties:

fern/apis/server/definition/score.yml

@@ -59,7 +59,7 @@ service:
             type: optional<commons.ScoreDataType>
             docs: Retrieve only scores with a specific dataType.
           traceTags:
-            type: optional<list<string>>
+            type: optional<string>
             allow-multiple: true
             docs: Only scores linked to traces that include all of these tags will be returned.
       response: GetScoresResponse

fern/apis/server/generators.yml

@@ -3,7 +3,7 @@ groups:
   local:
     generators:
       - name: fernapi/fern-openapi
-        version: 0.0.31
+        version: 0.1.7
         output:
           location: local-file-system
           path: ../../../web/public/generated/api

fern/fern.config.json

@@ -1,4 +1,5 @@
 {
-  "organization": "finto",
+  "organization": "langfuse",
   "version": "0.43.7"
-}
+}
+

package.json

@@ -1,6 +1,6 @@
 {
   "name": "langfuse",
-  "version": "3.2.0",
+  "version": "3.27.2",
   "author": "engineering@langfuse.com",
   "license": "MIT",
   "private": true,
@@ -25,7 +25,6 @@
     "dev": "turbo run dev",
     "lint": "turbo run lint",
     "test": "turbo run test",
-    "models:migrate": "turbo run models:migrate",
     "release": "dotenv -e ../.env -- release-it",
     "prepare": "husky"
   },
@@ -84,7 +83,11 @@
   "packageManager": "pnpm@9.5.0",
   "pnpm": {
     "overrides": {
-      "jsonpath-plus": "10.0.7"
+      "nanoid": "^3.3.8",
+      "katex": "^0.16.21"
+    },
+    "patchedDependencies": {
+      "next-auth@4.24.11": "patches/next-auth@4.24.11.patch"
     }
   }
 }

packages/shared/clickhouse/migrations/clustered/0007_add_event_log.down.sql

@@ -0,0 +1 @@
+DROP TABLE event_log ON CLUSTER default;

packages/shared/clickhouse/migrations/clustered/0007_add_event_log.up.sql

@@ -0,0 +1,19 @@
+CREATE TABLE event_log ON CLUSTER default
+(
+    `id`          String,
+    `project_id`  String,
+    `entity_type` String,
+    `entity_id`   String,
+    `event_id`    Nullable(String),
+
+    `bucket_name` String,
+    `bucket_path` String,
+
+    `created_at`  DateTime64(3) DEFAULT now(),
+    `updated_at`  DateTime64(3) DEFAULT now()
+) ENGINE = MergeTree()
+      ORDER BY (
+                project_id,
+                entity_type,
+                entity_id
+          );

packages/shared/clickhouse/migrations/unclustered/0007_add_event_log.down.sql

@@ -0,0 +1 @@
+DROP TABLE event_log;

packages/shared/clickhouse/migrations/unclustered/0007_add_event_log.up.sql

@@ -0,0 +1,19 @@
+CREATE TABLE event_log
+(
+    `id`          String,
+    `project_id`  String,
+    `entity_type` String,
+    `entity_id`   String,
+    `event_id`    Nullable(String),
+
+    `bucket_name` String,
+    `bucket_path` String,
+
+    `created_at`  DateTime64(3) DEFAULT now(),
+    `updated_at`  DateTime64(3) DEFAULT now()
+) ENGINE = MergeTree()
+      ORDER BY (
+                project_id,
+                entity_type,
+                entity_id
+          );

packages/shared/clickhouse/scripts/down.sh

@@ -18,21 +18,31 @@ then
     exit 1
 fi
 
+# Ensure CLICKHOUSE_DB is set
+if [ -z "${CLICKHOUSE_DB}" ]; then
+    export CLICKHOUSE_DB="default"
+fi
+
+# Ensure CLICKHOUSE_CLUSTER_NAME is set
+if [ -z "${CLICKHOUSE_CLUSTER_NAME}" ]; then
+    export CLICKHOUSE_CLUSTER_NAME="default"
+fi
+
 # Construct the database URL
 if [ "$CLICKHOUSE_CLUSTER_ENABLED" == "false" ] ; then
   if [ "$CLICKHOUSE_MIGRATION_SSL" = true ] ; then
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
   else
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&x-migrations-table-engine=MergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&x-migrations-table-engine=MergeTree"
   fi
 
   # Execute the up command
   migrate -source file://clickhouse/migrations/unclustered -database "$DATABASE_URL" down
 else
   if [ "$CLICKHOUSE_MIGRATION_SSL" = true ] ; then
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&secure=true&skip_verify=true&x-cluster-name=default&x-migrations-table-engine=ReplicatedMergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&secure=true&skip_verify=true&x-cluster-name=${CLICKHOUSE_CLUSTER_NAME}&x-migrations-table-engine=ReplicatedMergeTree"
   else
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&x-cluster-name=default&x-migrations-table-engine=ReplicatedMergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&x-cluster-name=${CLICKHOUSE_CLUSTER_NAME}&x-migrations-table-engine=ReplicatedMergeTree"
   fi
 
   # Execute the up command

packages/shared/clickhouse/scripts/drop.sh

@@ -12,11 +12,16 @@ then
     exit 1
 fi
 
+# Ensure CLICKHOUSE_DB is set
+if [ -z "${CLICKHOUSE_DB}" ]; then
+    export CLICKHOUSE_DB="default"
+fi
+
 # Construct the database URL
 if [ "$CLICKHOUSE_MIGRATION_SSL" = true ] ; then
-    DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
+    DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
 else
-    DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&x-migrations-table-engine=MergeTree"
+    DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&x-migrations-table-engine=MergeTree"
 fi
 # Execute the drop command
 migrate -source file://clickhouse/migrations -database "$DATABASE_URL" drop

packages/shared/clickhouse/scripts/up.sh

@@ -18,21 +18,31 @@ then
     exit 1
 fi
 
+# Ensure CLICKHOUSE_DB is set
+if [ -z "${CLICKHOUSE_DB}" ]; then
+    export CLICKHOUSE_DB="default"
+fi
+
+# Ensure CLICKHOUSE_CLUSTER_NAME is set
+if [ -z "${CLICKHOUSE_CLUSTER_NAME}" ]; then
+    export CLICKHOUSE_CLUSTER_NAME="default"
+fi
+
 # Construct the database URL
 if [ "$CLICKHOUSE_CLUSTER_ENABLED" == "false" ] ; then
   if [ "$CLICKHOUSE_MIGRATION_SSL" = true ] ; then
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&secure=true&skip_verify=true&x-migrations-table-engine=MergeTree"
   else
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&x-migrations-table-engine=MergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&x-migrations-table-engine=MergeTree"
   fi
 
   # Execute the up command
   migrate -source file://clickhouse/migrations/unclustered -database "$DATABASE_URL" up
 else
 if [ "$CLICKHOUSE_MIGRATION_SSL" = true ] ; then
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&secure=true&skip_verify=true&x-cluster-name=default&x-migrations-table-engine=ReplicatedMergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&secure=true&skip_verify=true&x-cluster-name=${CLICKHOUSE_CLUSTER_NAME}&x-migrations-table-engine=ReplicatedMergeTree"
   else
-      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=default&x-multi-statement=true&x-cluster-name=default&x-migrations-table-engine=ReplicatedMergeTree"
+      DATABASE_URL="${CLICKHOUSE_MIGRATION_URL}?username=${CLICKHOUSE_USER}&password=${CLICKHOUSE_PASSWORD}&database=${CLICKHOUSE_DB}&x-multi-statement=true&x-cluster-name=${CLICKHOUSE_CLUSTER_NAME}&x-migrations-table-engine=ReplicatedMergeTree"
   fi
 
   # Execute the up command

packages/shared/package.json

@@ -33,7 +33,7 @@
   "scripts": {
     "build": "tsc",
     "dev": "tsc --watch",
-    "lint": "eslint . --ext .js,.jsx,.ts,.tsx",
+    "lint": "eslint . --ext .js,.jsx,.ts,.tsx --max-warnings 110",
     "lint:fix": "eslint . --ext .js,.jsx,.ts,.tsx --fix",
     "db:migrate": "DISABLE_ERD=false dotenv -e ../../.env -- npx prisma migrate dev",
     "db:push": "DISABLE_ERD=false dotenv -e ../../.env -- npx prisma db push",
@@ -61,25 +61,26 @@
     "@aws-sdk/s3-request-presigner": "^3.679.0",
     "@azure/storage-blob": "^12.26.0",
     "@clickhouse/client": "^1.4.0",
-    "@langchain/anthropic": "^0.3.8",
-    "@langchain/aws": "^0.1.2",
-    "@langchain/core": "^0.3.18",
-    "@langchain/google-vertexai": "^0.1.3",
-    "@langchain/openai": "^0.3.14",
+    "@langchain/anthropic": "^0.3.12",
+    "@langchain/aws": "^0.1.3",
+    "@langchain/core": "^0.3.37",
+    "@langchain/google-vertexai": "^0.1.8",
+    "@langchain/openai": "^0.3.17",
     "@opentelemetry/api": ">=1.0.0 <1.10.0",
-    "@prisma/client": "^5.22.0",
+    "@prisma/client": "^6.3.0",
     "@react-email/components": "^0.0.19",
     "@react-email/render": "^0.0.15",
     "@types/bcryptjs": "^2.4.6",
     "axios": "^1.7.7",
     "bcryptjs": "^2.4.3",
-    "bullmq": "^5.12.10",
-    "dd-trace": "^5.23.1",
+    "bullmq": "^5.34.10",
+    "dd-trace": "^5.36.0",
     "decimal.js": "^10.4.3",
     "exponential-backoff": "^3.1.1",
+    "https-proxy-agent": "^7.0.6",
     "ioredis": "^5.4.1",
     "kysely": "^0.27.4",
-    "langchain": "^0.3.6",
+    "langchain": "^0.3.15",
     "langfuse-langchain": "3.30.3",
     "lodash": "^4.17.21",
     "next-auth": "^4.24.11",
@@ -106,7 +107,7 @@
     "kysely-codegen": "^0.16.8",
     "nodemon": "^3.1.7",
     "prettier": "^3.3.3",
-    "prisma": "^5.22.0",
+    "prisma": "^6.3.0",
     "prisma-erd-generator": "^1.11.2",
     "prisma-kysely": "^1.8.0",
     "ts-node": "^10.9.2",
@@ -121,7 +122,7 @@
   },
   "pnpm": {
     "overrides": {
-      "jsonpath-plus": "10.0.7"
+      "nanoid": "^3.3.8"
     }
   }
 }

packages/shared/prisma/generated/types.ts

@@ -170,6 +170,17 @@ export type BatchExport = {
     url: string | null;
     log: string | null;
 };
+export type BillingMeterBackup = {
+    stripe_customer_id: string;
+    meter_id: string;
+    start_time: Timestamp;
+    end_time: Timestamp;
+    aggregated_value: number;
+    event_name: string;
+    org_id: string;
+    created_at: Generated<Timestamp>;
+    updated_at: Generated<Timestamp>;
+};
 export type Comment = {
     id: string;
     project_id: string;
@@ -242,16 +253,6 @@ export type EvalTemplate = {
     vars: Generated<string[]>;
     output_schema: unknown;
 };
-export type Events = {
-    id: string;
-    created_at: Generated<Timestamp>;
-    updated_at: Generated<Timestamp>;
-    project_id: string;
-    data: unknown;
-    headers: Generated<unknown>;
-    url: string | null;
-    method: string | null;
-};
 export type JobConfiguration = {
     id: string;
     created_at: Generated<Timestamp>;
@@ -293,6 +294,8 @@ export type LlmApiKeys = {
     base_url: string | null;
     custom_models: Generated<string[]>;
     with_default_models: Generated<boolean>;
+    extra_headers: string | null;
+    extra_header_keys: Generated<string[]>;
     config: unknown | null;
     project_id: string;
 };
@@ -455,6 +458,7 @@ export type Project = {
     updated_at: Generated<Timestamp>;
     deleted_at: Timestamp | null;
     name: string;
+    retention_days: number | null;
 };
 export type ProjectMembership = {
     org_membership_id: string;
@@ -478,6 +482,14 @@ export type Prompt = {
     config: Generated<unknown>;
     tags: Generated<string[]>;
     labels: Generated<string[]>;
+    commit_message: string | null;
+};
+export type QueueBackUp = {
+    id: string;
+    project_id: string | null;
+    queue_name: string;
+    content: unknown;
+    created_at: Generated<Timestamp>;
 };
 export type Score = {
     id: string;
@@ -604,6 +616,7 @@ export type DB = {
     audit_logs: AuditLog;
     background_migrations: BackgroundMigration;
     batch_exports: BatchExport;
+    billing_meter_backups: BillingMeterBackup;
     comments: Comment;
     cron_jobs: CronJobs;
     dataset_items: DatasetItem;
@@ -611,7 +624,6 @@ export type DB = {
     dataset_runs: DatasetRuns;
     datasets: Dataset;
     eval_templates: EvalTemplate;
-    events: Events;
     job_configurations: JobConfiguration;
     job_executions: JobExecution;
     llm_api_keys: LlmApiKeys;
@@ -628,6 +640,7 @@ export type DB = {
     project_memberships: ProjectMembership;
     projects: Project;
     prompts: Prompt;
+    queue_backups: QueueBackUp;
     score_configs: ScoreConfig;
     scores: Score;
     Session: Session;

packages/shared/prisma/migrations/20250108220721_add_queue_backup_table/migration.sql

@@ -0,0 +1,13 @@
+-- CreateTable
+CREATE TABLE "queue_backups" (
+    "id" TEXT NOT NULL,
+    "project_id" TEXT,
+    "queue_name" TEXT NOT NULL,
+    "content" JSONB NOT NULL,
+    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+
+    CONSTRAINT "queue_backups_pkey" PRIMARY KEY ("id")
+);
+
+-- AddForeignKey
+ALTER TABLE "traces" ADD CONSTRAINT "traces_session_id_project_id_fkey" FOREIGN KEY ("session_id", "project_id") REFERENCES "trace_sessions"("id", "project_id") ON DELETE RESTRICT ON UPDATE CASCADE;

packages/shared/prisma/migrations/20250109083346_drop_trace_tracesession_fk/migration.sql

@@ -0,0 +1,2 @@
+-- DropForeignKey
+ALTER TABLE "traces" DROP CONSTRAINT "traces_session_id_project_id_fkey";

packages/shared/prisma/migrations/20250116154613_add_billing_meter_backups/migration.sql

@@ -0,0 +1,21 @@
+-- CreateTable
+CREATE TABLE "billing_meter_backups" (
+    "stripe_customer_id" TEXT NOT NULL,
+    "meter_id" TEXT NOT NULL,
+    "start_time" TIMESTAMP(3) NOT NULL,
+    "end_time" TIMESTAMP(3) NOT NULL,
+    
+    "aggregated_value" INTEGER NOT NULL,
+    
+    "event_name" TEXT NOT NULL,
+    "org_id" TEXT NOT NULL,
+    
+    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP
+);
+
+-- CreateIndex
+CREATE INDEX "billing_meter_backups_stripe_customer_id_meter_id_start_tim_idx" ON "billing_meter_backups"("stripe_customer_id", "meter_id", "start_time", "end_time");
+
+-- CreateIndex
+CREATE UNIQUE INDEX "billing_meter_backups_stripe_customer_id_meter_id_start_tim_key" ON "billing_meter_backups"("stripe_customer_id", "meter_id", "start_time", "end_time");

packages/shared/prisma/migrations/20250122152102_add_llm_api_keys_extra_headers/migration.sql

@@ -0,0 +1,3 @@
+ALTER TABLE "llm_api_keys"
+    ADD COLUMN "extra_headers" TEXT,
+    ADD COLUMN "extra_header_keys" TEXT[] NOT NULL DEFAULT '{}'::TEXT[];

packages/shared/prisma/migrations/20250123103200_add_retention_days_to_projects/migration.sql

@@ -0,0 +1,3 @@
+-- AlterTable
+ALTER TABLE "projects"
+    ADD COLUMN "retention_days" INTEGER;

packages/shared/prisma/migrations/20250128144418_llm_adapter_rename_google_vertex_ai/migration.sql

@@ -0,0 +1 @@
+UPDATE "llm_api_keys" SET adapter = 'google-vertex-ai' WHERE adapter = 'vertex-ai';

packages/shared/prisma/migrations/20250128163035_add_nullable_commit_message_prompts/migration.sql

@@ -0,0 +1,2 @@
+-- AlterTable
+ALTER TABLE "prompts" ADD COLUMN     "commit_message" TEXT;

packages/shared/prisma/migrations/20250204180200_add_event_log_table/migration.sql

@@ -0,0 +1,20 @@
+-- CreateTable
+CREATE TABLE "event_log" (
+    "id" TEXT NOT NULL,
+    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updated_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "bucket_name" TEXT NOT NULL,
+    "bucket_path" TEXT NOT NULL,
+    "project_id" TEXT NOT NULL,
+    "entity_type" TEXT NOT NULL,
+    "entity_id" TEXT NOT NULL,
+    "event_id" TEXT NOT NULL,
+
+    "trace_id" TEXT,
+
+    CONSTRAINT "event_log_pkey" PRIMARY KEY ("id"),
+    CONSTRAINT "event_log_project_id_fkey" FOREIGN KEY ("project_id") REFERENCES "projects"("id") ON DELETE CASCADE ON UPDATE CASCADE
+);
+
+-- CreateIndex
+CREATE INDEX "event_log_project_id_entity_type_entity_id_idx" ON "event_log"("project_id", "entity_type", "entity_id");

packages/shared/prisma/migrations/20250211102600_drop_event_log_table/migration.sql

@@ -0,0 +1,2 @@
+-- DropTable
+DROP TABLE event_log;

packages/shared/prisma/migrations/20250211123300_drop_events_table/migration.sql

@@ -0,0 +1,2 @@
+-- DropTable
+DROP TABLE events;

packages/shared/prisma/schema.prisma

@@ -115,13 +115,13 @@ model Project {
   updatedAt           DateTime               @default(now()) @updatedAt @map("updated_at")
   deletedAt           DateTime?              @map("deleted_at")
   name                String
+  retentionDays       Int?                   @map("retention_days")
   projectMembers      ProjectMembership[]
   organization        Organization           @relation(fields: [orgId], references: [id], onUpdate: Cascade, onDelete: Cascade)
   traces              Trace[]
   observations        Observation[]
   apiKeys             ApiKey[]
   dataset             Dataset[]
-  RawEvents           Events[]
   invitations         MembershipInvitation[]
   sessions            TraceSession[]
   Prompt              Prompt[]
@@ -192,6 +192,8 @@ model LlmApiKeys {
   baseURL           String?  @map("base_url")
   customModels      String[] @default([]) @map("custom_models")
   withDefaultModels Boolean  @default(true) @map("with_default_models")
+  extraHeaders      String?  @map("extra_headers")
+  extraHeaderKeys   String[] @default([]) @map("extra_header_keys")
   config            Json?
 
   projectId String  @map("project_id")
@@ -272,7 +274,6 @@ model TraceSession {
   project    Project  @relation(fields: [projectId], references: [id], onDelete: Cascade)
   bookmarked Boolean  @default(false)
   public     Boolean  @default(false)
-  traces     Trace[]
 
   @@id([id, projectId])
   @@index([projectId])
@@ -284,25 +285,24 @@ model TraceSession {
 // Update TraceView below when making changes to this model!
 
 model Trace {
-  id         String        @id @default(cuid())
-  externalId String?       @map("external_id")
-  timestamp  DateTime      @default(now())
+  id         String   @id @default(cuid())
+  externalId String?  @map("external_id")
+  timestamp  DateTime @default(now())
   name       String?
-  userId     String?       @map("user_id")
+  userId     String?  @map("user_id")
   metadata   Json?
   release    String?
   version    String?
-  projectId  String        @map("project_id")
-  project    Project       @relation(fields: [projectId], references: [id], onDelete: Cascade)
-  public     Boolean       @default(false)
-  bookmarked Boolean       @default(false)
-  tags       String[]      @default([])
+  projectId  String   @map("project_id")
+  project    Project  @relation(fields: [projectId], references: [id], onDelete: Cascade)
+  public     Boolean  @default(false)
+  bookmarked Boolean  @default(false)
+  tags       String[] @default([])
   input      Json?
   output     Json?
-  sessionId  String?       @map("session_id")
-  session    TraceSession? @relation(fields: [sessionId, projectId], references: [id, projectId])
-  createdAt  DateTime      @default(now()) @map("created_at")
-  updatedAt  DateTime      @default(now()) @updatedAt @map("updated_at")
+  sessionId  String?  @map("session_id")
+  createdAt  DateTime @default(now()) @map("created_at")
+  updatedAt  DateTime @default(now()) @updatedAt @map("updated_at")
 
   @@index([projectId, timestamp])
   @@index([sessionId])
@@ -690,21 +690,6 @@ model DatasetRunItems {
   @@map("dataset_run_items")
 }
 
-model Events {
-  id        String   @id @default(cuid())
-  createdAt DateTime @default(now()) @map("created_at")
-  updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
-  projectId String   @map("project_id")
-  project   Project  @relation(fields: [projectId], references: [id], onDelete: Cascade)
-  data      Json
-  headers   Json     @default("{}")
-  url       String?
-  method    String?
-
-  @@index(projectId)
-  @@map("events")
-}
-
 model Comment {
   id           String            @id @default(cuid())
   projectId    String            @map("project_id")
@@ -737,14 +722,15 @@ model Prompt {
 
   createdBy String @map("created_by")
 
-  prompt   Json
-  name     String
-  version  Int
-  type     String   @default("text")
-  isActive Boolean? @map("is_active") // Deprecated. To be removed once 'production' labels work as expected.
-  config   Json     @default("{}") @db.Json
-  tags     String[] @default([])
-  labels   String[] @default([])
+  prompt        Json
+  name          String
+  version       Int
+  type          String   @default("text")
+  isActive      Boolean? @map("is_active") // Deprecated. To be removed once 'production' labels work as expected.
+  config        Json     @default("{}") @db.Json
+  tags          String[] @default([])
+  labels        String[] @default([])
+  commitMessage String?  @map("commit_message")
 
   @@unique([projectId, name, version])
   @@index([projectId, id])
@@ -1017,3 +1003,37 @@ model ObservationMedia {
   @@index([projectId, observationId])
   @@map("observation_media")
 }
+
+model QueueBackUp {
+  id        String  @id @default(cuid())
+  projectId String? @map("project_id")
+  queueName String  @map("queue_name")
+  content   Json
+
+  createdAt DateTime @default(now()) @map("created_at")
+
+  @@map("queue_backups")
+}
+
+model BillingMeterBackup {
+  // unique
+  stripeCustomerId String   @map("stripe_customer_id")
+  meterId          String   @map("meter_id")
+  startTime        DateTime @map("start_time")
+  endTime          DateTime @map("end_time")
+
+  // value
+  aggregatedValue Int @map("aggregated_value")
+
+  // labels
+  eventName String @map("event_name")
+  orgId     String @map("org_id")
+
+  // ts
+  createdAt DateTime @default(now()) @map("created_at")
+  updatedAt DateTime @default(now()) @updatedAt @map("updated_at")
+
+  @@unique([stripeCustomerId, meterId, startTime, endTime])
+  @@index([stripeCustomerId, meterId, startTime, endTime])
+  @@map("billing_meter_backups")
+}

packages/shared/scripts/prepareClickhouse.ts

@@ -97,19 +97,31 @@ export const prepareClickhouse = async (
       repeat('input', toInt64(randExponential(1 / 100))) AS input,
       repeat('output', toInt64(randExponential(1 / 100))) AS output,
       case
-        when number % 2 = 0 then 'claude-3-haiku-20240307'
+        when number % 2 = 0 then 'clause-3-haiku-20230407'
         else 'gpt-4'
       end as provided_model_name,
       case
-        when number % 2 = 0 then 'cltr0w45b000008k1407o9qv1'
-        else 'clrntkjgy000f08jx79v9g1xj'
+        when number % 2 = 0 then 'cltra4wbs0000k1407g0ya3'
+        else '1cmtk9y0000y3y79x9jgxj'
       end as internal_model_id,
-      '{"temperature": 0.7, "max_tokens": 150}' AS model_parameters,
-      map('input', toUInt64(randUniform(0, 1000)), 'output', toUInt64(randUniform(0, 1000)), 'total', toUInt64(randUniform(0, 2000))) AS provided_usage_details,
-      map('input', toUInt64(randUniform(0, 1000)), 'output', toUInt64(randUniform(0, 1000)), 'total', toUInt64(randUniform(0, 2000))) AS usage_details,
-      map('input', toDecimal64(randUniform(0, 1000), 12), 'output', toDecimal64(randUniform(0, 1000), 12), 'total', toDecimal64(randUniform(0, 2000), 12)) AS provided_cost_details,
-      map('input', toDecimal64(randUniform(0, 1000), 12), 'output', toDecimal64(randUniform(0, 1000), 12), 'total', toDecimal64(randUniform(0, 2000), 12)) AS cost_details,
-      toDecimal64(randUniform(0, 2000), 12) AS total_cost,
+      if("type" = 'GENERATION',
+        '{"temperature": 0.7, "max_tokens": 150}',
+        '{}') AS model_parameters,
+      if("type" = 'GENERATION',
+        map('input', toUInt64(randUniform(0, 1000)), 'output', toUInt64(randUniform(0, 1000)), 'total', toUInt64(randUniform(0, 2000))),
+        map()) AS provided_usage_details,
+      if("type" = 'GENERATION',
+        map('input', toUInt64(randUniform(0, 1000)), 'output', toUInt64(randUniform(0, 1000)), 'total', toUInt64(randUniform(0, 2000))),
+        map()) AS usage_details,
+      if("type" = 'GENERATION',
+        map('input', toDecimal64(randUniform(0, 1000), 12), 'output', toDecimal64(randUniform(0, 1000), 12), 'total', toDecimal64(randUniform(0, 2000), 12)),
+        map()) AS provided_cost_details,
+      if("type" = 'GENERATION',
+        map('input', toDecimal64(randUniform(0, 1000), 12), 'output', toDecimal64(randUniform(0, 1000), 12), 'total', toDecimal64(randUniform(0, 2000), 12)),
+        map()) AS cost_details,
+      if("type" = 'GENERATION',
+        toDecimal64(randUniform(0, 2000), 12),
+        NULL) AS total_cost,
       addMilliseconds(start_time, if(rand() < 0.6, floor(randUniform(0, 500)), floor(randUniform(0, 600)))) AS completion_start_time,
       array(${SEED_PROMPTS.map((p) => `concat('${p.id}',project_id)`).join(
         ",",

packages/shared/src/env.ts

@@ -29,17 +29,10 @@ const EnvSchema = z.object({
   LANGFUSE_CACHE_PROMPT_ENABLED: z.enum(["true", "false"]).default("false"),
   LANGFUSE_CACHE_PROMPT_TTL_SECONDS: z.coerce.number().default(60 * 60),
   CLICKHOUSE_URL: z.string().url(),
+  CLICKHOUSE_CLUSTER_NAME: z.string().default("default"),
+  CLICKHOUSE_DB: z.string().default("default"),
   CLICKHOUSE_USER: z.string(),
   CLICKHOUSE_PASSWORD: z.string(),
-  LANGFUSE_SDK_CI_SYNC_PROCESSING_ENABLED: z
-    .enum(["true", "false"])
-    .default("false"),
-  LANGFUSE_CLICKHOUSE_INGESTION_ENABLED: z
-    .enum(["true", "false"])
-    .default("true"),
-  LANGFUSE_POSTGRES_INGESTION_ENABLED: z
-    .enum(["true", "false"])
-    .default("false"),
 
   LANGFUSE_INGESTION_QUEUE_DELAY_MS: z.coerce
     .number()
@@ -66,6 +59,13 @@ const EnvSchema = z.object({
     .default("false"),
   LANGFUSE_USE_AZURE_BLOB: z.enum(["true", "false"]).default("false"),
   STRIPE_SECRET_KEY: z.string().optional(),
+
+  LANGFUSE_S3_CORE_DATA_EXPORT_IS_ENABLED: z
+    .enum(["true", "false"])
+    .default("false"),
+  LANGFUSE_POSTGRES_METERING_DATA_EXPORT_IS_ENABLED: z
+    .enum(["true", "false"])
+    .default("false"),
 });
 
 export const env: z.infer<typeof EnvSchema> =

packages/shared/src/errors/ApiError.ts

@@ -1,7 +1,7 @@
 import { BaseError } from "./BaseError";
 
 export class ApiError extends BaseError {
-  constructor(description = "Api call failed") {
-    super("ApiError", 500, description, true);
+  constructor(description = "Api call failed", status = 500) {
+    super("ApiError", status, description, true);
   }
 }

packages/shared/src/features/batchAction/types.ts

@@ -0,0 +1,38 @@
+import z from "zod";
+import { singleFilter } from "../../interfaces/filters";
+import { orderBy } from "../../interfaces/orderBy";
+import { BatchExportTableName } from "../batchExport/types";
+
+/* eslint-disable no-unused-vars */
+export enum BatchActionType {
+  Create = "create",
+  Delete = "delete",
+}
+
+const ActionIdSchema = z.enum([
+  "trace-delete",
+  "trace-add-to-annotation-queue",
+]);
+
+export type ActionId = z.infer<typeof ActionIdSchema>;
+
+export const BatchActionQuerySchema = z.object({
+  filter: z.array(singleFilter).nullable(),
+  orderBy,
+});
+
+export type BatchActionQuery = z.infer<typeof BatchActionQuerySchema>;
+
+export const CreateBatchActionSchema = z.object({
+  projectId: z.string(),
+  actionId: ActionIdSchema,
+  targetId: z.string().optional(),
+  query: BatchActionQuerySchema,
+  tableName: z.nativeEnum(BatchExportTableName),
+});
+
+export const GetIsBatchActionInProgressSchema = z.object({
+  projectId: z.string(),
+  actionId: ActionIdSchema,
+  tableName: z.nativeEnum(BatchExportTableName),
+});

packages/shared/src/features/evals/types.ts

@@ -17,6 +17,7 @@ export const variableMapping = z
     objectName: z.string().nullish(),
     langfuseObject: z.enum(langfuseObjects),
     selectedColumnId: z.string(),
+    jsonSelector: z.string().nullish(),
   })
   .refine(
     (value) => value.langfuseObject === "trace" || value.objectName !== null,
@@ -32,6 +33,7 @@ export const wipVariableMapping = z.object({
   objectName: z.string().nullish(),
   langfuseObject: z.enum(langfuseObjects),
   selectedColumnId: z.string().nullish(),
+  jsonSelector: z.string().nullish(),
 });
 
 const observationCols = [

packages/shared/src/index.ts

@@ -21,7 +21,10 @@ export * from "./server/llm/types";
 
 // evals
 export * from "./features/evals/types";
+
+// table actions
 export * from "./features/batchExport/types";
+export * from "./features/batchAction/types";
 
 // annotation
 export * from "./features/annotation/types";

packages/shared/src/observationsTable.ts

@@ -20,6 +20,13 @@ export const observationsTableCols: ColumnDefinition[] = [
     options: [], // to be added at runtime
     nullable: true,
   },
+  {
+    name: "type",
+    id: "type",
+    type: "stringOptions",
+    options: [],
+    internal: 'o."type"',
+  },
   { name: "Trace ID", id: "traceId", type: "string", internal: 't."id"' },
   {
     name: "Trace Name",
@@ -112,6 +119,14 @@ export const observationsTableCols: ColumnDefinition[] = [
     options: [], // to be added at runtime
     nullable: true,
   },
+  {
+    name: "Model ID",
+    id: "modelId",
+    type: "stringOptions",
+    internal: 'o."internal_model_id"',
+    options: [], // to be added at runtime
+    nullable: true,
+  },
   {
     name: "Input Tokens",
     id: "inputTokens",
@@ -186,20 +201,25 @@ export const observationsTableCols: ColumnDefinition[] = [
 // allows for undefined options, to offer filters while options are still loading
 export type ObservationOptions = {
   model: Array<OptionsDefinition>;
+  modelId: Array<OptionsDefinition>;
   name: Array<OptionsDefinition>;
   traceName: Array<OptionsDefinition>;
   scores_avg: Array<string>;
   promptName: Array<OptionsDefinition>;
   tags: Array<OptionsDefinition>;
+  type: Array<OptionsDefinition>;
 };
 
 export function observationsTableColsWithOptions(
-  options?: ObservationOptions
+  options?: ObservationOptions,
 ): ColumnDefinition[] {
   return observationsTableCols.map((col) => {
     if (col.id === "model") {
       return { ...col, options: options?.model ?? [] };
     }
+    if (col.id === "modelId") {
+      return { ...col, options: options?.modelId ?? [] };
+    }
     if (col.id === "name") {
       return { ...col, options: options?.name ?? [] };
     }
@@ -215,6 +235,9 @@ export function observationsTableColsWithOptions(
     if (col.id === "tags") {
       return { ...col, options: options?.tags ?? [] };
     }
+    if (col.id === "type") {
+      return { ...col, options: options?.type ?? [] };
+    }
     return col;
   });
 }

packages/shared/src/server/auth/customSsoProvider.ts

@@ -17,7 +17,6 @@ export function CustomSSOProvider<P extends CustomSSOUser>(
     wellKnown: `${options.issuer}/.well-known/openid-configuration`,
     authorization: { params: { scope: "openid email profile" } }, // overridden by options.authorization to be able to set custom scopes, deep merged with this default
     checks: ["pkce", "state"],
-    idToken: true,
     profile(profile) {
       return {
         id: profile.sub,

packages/shared/src/server/auth/gitHubEnterpriseProvider.ts

@@ -2,64 +2,64 @@ import type { OAuthConfig, OAuthUserConfig } from "next-auth/providers/oauth";
 import type { GithubProfile, GithubEmail } from "next-auth/providers/github";
 
 export function GitHubEnterpriseProvider<P extends GithubProfile>(
-    options: OAuthUserConfig<P> & {
-        enterprise?: {
-            baseUrl?: string;
-        };
-    }
+  options: OAuthUserConfig<P> & {
+    enterprise?: {
+      baseUrl?: string;
+    };
+  },
 ): OAuthConfig<P> {
-    const baseUrl = options?.enterprise?.baseUrl ?? "https://github.com"
-    const apiBaseUrl = options?.enterprise?.baseUrl
-        ? `${options?.enterprise?.baseUrl}/api/v3`
-        : "https://api.github.com"
+  const baseUrl = options?.enterprise?.baseUrl ?? "https://github.com";
+  const apiBaseUrl = options?.enterprise?.baseUrl
+    ? `${options?.enterprise?.baseUrl}/api/v3`
+    : "https://api.github.com";
 
-    return {
-        id: "github-enterprise",
-        name: "GitHub Enterprise",
-        type: "oauth",
-        authorization: {
-            url: `${baseUrl}/login/oauth/authorize`,
-            params: { scope: "read:user user:email" },
-        },
-        token: `${baseUrl}/login/oauth/access_token`,
-        userinfo: {
-            url: `${apiBaseUrl}/user`,
-            async request({ client, tokens }) {
-                const profile = await client.userinfo(tokens.access_token!)
+  return {
+    id: "github-enterprise",
+    name: "GitHub Enterprise",
+    type: "oauth",
+    authorization: {
+      url: `${baseUrl}/login/oauth/authorize`,
+      params: { scope: "read:user user:email" },
+    },
+    token: `${baseUrl}/login/oauth/access_token`,
+    userinfo: {
+      url: `${apiBaseUrl}/user`,
+      async request({ client, tokens }) {
+        const profile = await client.userinfo(tokens.access_token!);
 
-                if (!profile.email) {
-                    // If the user does not have a public email, get another via the GitHub API
-                    // See https://docs.github.com/en/rest/users/emails#list-email-addresses-for-the-authenticated-user
-                    const res = await fetch(`${apiBaseUrl}/user/emails`, {
-                        headers: { Authorization: `token ${tokens.access_token}` },
-                    })
+        if (!profile.email) {
+          // If the user does not have a public email, get another via the GitHub API
+          // See https://docs.github.com/en/rest/users/emails#list-email-addresses-for-the-authenticated-user
+          const res = await fetch(`${apiBaseUrl}/user/emails`, {
+            headers: { Authorization: `token ${tokens.access_token}` },
+          });
 
-                    if (res.ok) {
-                        const emails: GithubEmail[] = await res.json()
-                        profile.email = (emails.find((e) => e.primary) ?? emails[0]).email
-                    }
-                }
+          if (res.ok) {
+            const emails = (await res.json()) as GithubEmail[];
+            profile.email = (emails.find((e) => e.primary) ?? emails[0]).email;
+          }
+        }
 
-                return profile
-            },
-        },
-        profile(profile) {
-            return {
-                id: profile.id.toString(),
-                name: profile.name ?? profile.login,
-                email: profile.email,
-                image: profile.avatar_url,
-            }
-        },
-        style: {
-            logo: "https://raw.githubusercontent.com/nextauthjs/next-auth/main/packages/next-auth/provider-logos/github.svg",
-            logoDark:
-                "https://raw.githubusercontent.com/nextauthjs/next-auth/main/packages/next-auth/provider-logos/github-dark.svg",
-            bg: "#fff",
-            bgDark: "#000",
-            text: "#000",
-            textDark: "#fff",
-        },
-        options,
-    }
-}
+        return profile;
+      },
+    },
+    profile(profile) {
+      return {
+        id: profile.id.toString(),
+        name: profile.name ?? profile.login,
+        email: profile.email,
+        image: profile.avatar_url,
+      };
+    },
+    style: {
+      logo: "https://raw.githubusercontent.com/nextauthjs/next-auth/main/packages/next-auth/provider-logos/github.svg",
+      logoDark:
+        "https://raw.githubusercontent.com/nextauthjs/next-auth/main/packages/next-auth/provider-logos/github-dark.svg",
+      bg: "#fff",
+      bgDark: "#000",
+      text: "#000",
+      textDark: "#fff",
+    },
+    options,
+  };
+}

packages/shared/src/server/clickhouse/client.ts

@@ -1,27 +1,34 @@
 import { createClient } from "@clickhouse/client";
 import { env } from "../../env";
 import { NodeClickHouseClientConfigOptions } from "@clickhouse/client/dist/config";
+import { getCurrentSpan } from "../instrumentation";
+import { propagation, context } from "@opentelemetry/api";
 
 export type ClickhouseClientType = ReturnType<typeof createClient>;
 
-export const clickhouseClient = (opts?: NodeClickHouseClientConfigOptions) =>
-  createClient({
+export const clickhouseClient = (opts?: NodeClickHouseClientConfigOptions) => {
+  const headers = opts?.http_headers ?? {};
+  const activeSpan = getCurrentSpan();
+  if (activeSpan) {
+    propagation.inject(context.active(), headers);
+  }
+  return createClient({
     ...opts,
     url: env.CLICKHOUSE_URL,
     username: env.CLICKHOUSE_USER,
     password: env.CLICKHOUSE_PASSWORD,
-    database: "default",
+    database: env.CLICKHOUSE_DB,
+    http_headers: headers,
     clickhouse_settings: {
       async_insert: 1,
       wait_for_async_insert: 1, // if disabled, we won't get errors from clickhouse
     },
   });
+};
 
-export const defaultClickhouseClient = clickhouseClient();
 /**
  * Accepts a JavaScript date and returns the DateTime in format YYYY-MM-DD HH:MM:SS
  */
-
 export const convertDateToClickhouseDateTime = (date: Date): string => {
   // 2024-11-06T20:37:00.123Z -> 2024-11-06 21:37:00.123
   return date.toISOString().replace("T", " ").replace("Z", "");

packages/shared/src/server/index.ts

@@ -8,6 +8,7 @@ export * from "./auth/apiKeys";
 export * from "./auth/customSsoProvider";
 export * from "./auth/gitHubEnterpriseProvider";
 export * from "./llm/fetchLLMCompletion";
+export * from "./llm/utils";
 export * from "./llm/types";
 export * from "./utils/DatabaseReadStream";
 export * from "./utils/transforms";
@@ -28,15 +29,17 @@ export * from "./redis/traceDelete";
 export * from "./redis/projectDelete";
 export * from "./redis/datasetRunItemUpsert";
 export * from "./redis/batchExport";
-export * from "./redis/legacyIngestion";
+export * from "./redis/batchActionQueue";
 export * from "./redis/ingestionQueue";
 export * from "./redis/postHogIntegrationQueue";
 export * from "./redis/postHogIntegrationProcessingQueue";
+export * from "./redis/dataRetentionQueue";
+export * from "./redis/dataRetentionProcessingQueue";
+export * from "./redis/coreDataS3ExportQueue";
+export * from "./redis/meteringDataPostgresExportQueue";
 export * from "./redis/experimentCreateQueue";
 export * from "./auth/types";
-export * from "./ingestion/legacy/index";
 export * from "./queues";
-export * from "./ingestion/legacy/EventProcessor";
 export * from "./orderByToPrisma";
 export * from "./filterToPrisma";
 export * from "./instrumentation";
@@ -44,6 +47,7 @@ export * from "./logger";
 export * from "./queries";
 export * from "./repositories";
 export * from "./redis/evalExecutionQueue";
+export * from "./services/sessions-ui-table-service";
 
 // test utils
 export * from "./test-utils";

packages/shared/src/server/ingestion/legacy/EventProcessor.ts

@@ -1,796 +0,0 @@
-import { v4 } from "uuid";
-import { type z } from "zod";
-import Decimal from "decimal.js";
-import { findModel } from "../modelMatch";
-import {
-  ObservationEvent,
-  eventTypes,
-  legacyObservationCreateEvent,
-  generationCreateEvent,
-  traceEvent,
-  scoreEvent,
-  sdkLogEvent,
-  ingestionEvent,
-} from "../types";
-import { validateAndInflateScore } from "../validateAndInflateScore";
-import { Trace, Observation, Score, Prisma, Model } from "@prisma/client";
-import { ForbiddenError, LangfuseNotFoundError } from "../../../errors";
-import { mergeJson } from "../../../utils/json";
-import { jsonSchema } from "../../../utils/zod";
-import { prisma } from "../../../db";
-import { LegacyIngestionAccessScope } from ".";
-import { logger } from "../../logger";
-import { env } from "../../../env";
-import { upsertTrace } from "../../repositories";
-import { convertDateToClickhouseDateTime } from "../../clickhouse/client";
-
-export interface EventProcessor {
-  auth(apiScope: LegacyIngestionAccessScope): void;
-
-  process(
-    apiScope: LegacyIngestionAccessScope,
-  ): Promise<Trace | Observation | Score> | undefined;
-}
-
-export const getProcessorForEvent = (
-  event: z.infer<typeof ingestionEvent>,
-  calculateTokenDelegate: (p: {
-    model: Model;
-    text: unknown;
-  }) => number | undefined,
-): EventProcessor => {
-  switch (event.type) {
-    case eventTypes.TRACE_CREATE:
-      return new TraceProcessor(event);
-    case eventTypes.OBSERVATION_CREATE:
-    case eventTypes.OBSERVATION_UPDATE:
-    case eventTypes.EVENT_CREATE:
-    case eventTypes.SPAN_CREATE:
-    case eventTypes.SPAN_UPDATE:
-    case eventTypes.GENERATION_CREATE:
-    case eventTypes.GENERATION_UPDATE:
-      return new ObservationProcessor(event, calculateTokenDelegate);
-    case eventTypes.SCORE_CREATE: {
-      return new ScoreProcessor(event);
-    }
-    case eventTypes.SDK_LOG:
-      return new SdkLogProcessor(event);
-  }
-};
-
-export class ObservationProcessor implements EventProcessor {
-  event: ObservationEvent;
-  calculateTokenDelegate: (p: {
-    model: Model;
-    text: unknown;
-  }) => number | undefined;
-
-  constructor(
-    event: ObservationEvent,
-    calculateTokenDelegate: (p: {
-      model: Model;
-      text: unknown;
-    }) => number | undefined,
-  ) {
-    this.event = event;
-    this.calculateTokenDelegate = calculateTokenDelegate;
-  }
-
-  async convertToObservation(
-    apiScope: LegacyIngestionAccessScope,
-    existingObservation: Omit<Observation, "input" | "output"> | null,
-  ): Promise<{
-    id: string;
-    create: Prisma.ObservationUncheckedCreateInput;
-    update: Prisma.ObservationUncheckedUpdateInput;
-  }> {
-    let type: "EVENT" | "SPAN" | "GENERATION";
-    switch (this.event.type) {
-      case eventTypes.OBSERVATION_CREATE:
-      case eventTypes.OBSERVATION_UPDATE:
-        type = this.event.body.type;
-        break;
-      case eventTypes.EVENT_CREATE:
-        type = "EVENT" as const;
-        break;
-      case eventTypes.SPAN_CREATE:
-      case eventTypes.SPAN_UPDATE:
-        type = "SPAN" as const;
-        break;
-      case eventTypes.GENERATION_CREATE:
-      case eventTypes.GENERATION_UPDATE:
-        type = "GENERATION" as const;
-        break;
-    }
-
-    if (
-      this.event.type === eventTypes.OBSERVATION_UPDATE &&
-      !existingObservation
-    ) {
-      throw new LangfuseNotFoundError(
-        `Observation with id ${this.event.id} not found`,
-      );
-    }
-
-    // find matching model definition based on event and existing observation in db
-    const internalModel: Model | undefined | null =
-      type === "GENERATION"
-        ? await findModel({
-            event: {
-              projectId: apiScope.projectId,
-              model:
-                "model" in this.event.body
-                  ? (this.event.body.model ?? undefined)
-                  : undefined,
-              unit:
-                "usage" in this.event.body
-                  ? (this.event.body.usage?.unit ?? undefined)
-                  : undefined,
-              startTime: this.event.body.startTime
-                ? new Date(this.event.body.startTime)
-                : undefined,
-            },
-            existingDbObservation: existingObservation ?? undefined,
-          })
-        : undefined;
-
-    // Token counts
-    const [newInputCount, newOutputCount] =
-      "usage" in this.event.body || "usageDetails" in this.event.body
-        ? await this.calculateTokenCounts(
-            apiScope.projectId,
-            this.event.body,
-            this.calculateTokenDelegate,
-            internalModel ?? undefined,
-            existingObservation ?? undefined,
-          )
-        : [undefined, undefined];
-
-    const newTotalCount =
-      "usage" in this.event.body || "usageDetails" in this.event.body
-        ? (this.event.body.usage?.total ??
-          this.event.body.usageDetails?.total ??
-          (newInputCount != null || newOutputCount != null
-            ? (newInputCount ?? 0) + (newOutputCount ?? 0)
-            : undefined))
-        : undefined;
-
-    const userProvidedTokenCosts = {
-      inputCost:
-        "usage" in this.event.body && this.event.body.usage?.inputCost != null // inputCost can be explicitly 0. Note only one equal sign to capture null AND undefined
-          ? new Decimal(this.event.body.usage?.inputCost)
-          : "costDetails" in this.event.body &&
-              this.event.body.costDetails?.input != null
-            ? new Decimal(this.event.body.costDetails?.input)
-            : existingObservation?.inputCost,
-      outputCost:
-        "usage" in this.event.body && this.event.body.usage?.outputCost != null // outputCost can be explicitly 0. Note only one equal sign to capture null AND undefined
-          ? new Decimal(this.event.body.usage?.outputCost)
-          : "costDetails" in this.event.body &&
-              this.event.body.costDetails?.output != null
-            ? new Decimal(this.event.body.costDetails?.output)
-            : existingObservation?.outputCost,
-      totalCost:
-        "usage" in this.event.body && this.event.body.usage?.totalCost != null // totalCost can be explicitly 0. Note only one equal sign to capture null AND undefined
-          ? new Decimal(this.event.body.usage?.totalCost)
-          : "costDetails" in this.event.body &&
-              this.event.body.costDetails?.total != null
-            ? new Decimal(this.event.body.costDetails?.total)
-            : existingObservation?.totalCost,
-    };
-
-    const tokenCounts = {
-      input: newInputCount ?? existingObservation?.promptTokens,
-      output: newOutputCount ?? existingObservation?.completionTokens,
-      total: newTotalCount || existingObservation?.totalTokens,
-    };
-
-    const calculatedCosts = ObservationProcessor.calculateTokenCosts(
-      internalModel,
-      userProvidedTokenCosts,
-      tokenCounts,
-    );
-
-    // merge metadata from existingObservation.metadata and metadata
-    const mergedMetadata = mergeJson(
-      existingObservation?.metadata
-        ? jsonSchema.parse(existingObservation.metadata)
-        : undefined,
-      this.event.body.metadata ?? undefined,
-    );
-
-    const prompt =
-      "promptName" in this.event.body &&
-      typeof this.event.body.promptName === "string" &&
-      "promptVersion" in this.event.body &&
-      typeof this.event.body.promptVersion === "number"
-        ? await prisma.prompt.findUnique({
-            where: {
-              projectId_name_version: {
-                projectId: apiScope.projectId,
-                name: this.event.body.promptName,
-                version: this.event.body.promptVersion,
-              },
-            },
-          })
-        : undefined;
-
-    // Only null if promptName and promptVersion are set but prompt is not found
-    if (prompt === null) {
-      logger.warn("Prompt not found for observation", this.event.body);
-    }
-
-    const observationId =
-      this.event.body.id ??
-      (() => {
-        const newId = v4();
-        logger.info(
-          `observation.id is null. Generating for projectId: ${apiScope.projectId}, id: ${newId}`,
-        );
-        return newId;
-      })();
-
-    let traceId = this.event.body?.traceId;
-    if (!this.event.body.traceId && !existingObservation) {
-      // Create trace if no traceId
-      traceId = observationId;
-
-      // Insert trace into postgres
-      await prisma.trace.upsert({
-        where: {
-          id: observationId,
-        },
-        create: {
-          projectId: apiScope.projectId,
-          name: this.event.body.name,
-          id: observationId,
-          timestamp: this.event.body.startTime || new Date(),
-        },
-        update: {},
-      });
-
-      if (env.CLICKHOUSE_URL) {
-        // Insert trace into clickhouse if enabled
-        await upsertTrace({
-          id: observationId,
-          project_id: apiScope.projectId,
-          timestamp: convertDateToClickhouseDateTime(
-            this.event.body.startTime
-              ? new Date(this.event.body.startTime)
-              : new Date(),
-          ),
-          created_at: convertDateToClickhouseDateTime(new Date()),
-          updated_at: convertDateToClickhouseDateTime(new Date()),
-        });
-      }
-    }
-
-    return {
-      id: observationId,
-      create: {
-        id: observationId,
-        traceId,
-        type: type,
-        name: this.event.body.name,
-        startTime: this.event.body.startTime
-          ? new Date(this.event.body.startTime)
-          : undefined,
-        endTime:
-          "endTime" in this.event.body && this.event.body.endTime
-            ? new Date(this.event.body.endTime)
-            : undefined,
-        completionStartTime:
-          "completionStartTime" in this.event.body &&
-          this.event.body.completionStartTime
-            ? new Date(this.event.body.completionStartTime)
-            : undefined,
-        metadata: mergedMetadata ?? this.event.body.metadata ?? undefined,
-        model: "model" in this.event.body ? this.event.body.model : undefined,
-        modelParameters:
-          "modelParameters" in this.event.body
-            ? (this.event.body.modelParameters ?? undefined)
-            : undefined,
-        input: this.event.body.input ?? undefined,
-        output: this.event.body.output ?? undefined,
-        promptTokens: newInputCount,
-        completionTokens: newOutputCount,
-        totalTokens: newTotalCount,
-        unit:
-          "usage" in this.event.body
-            ? (this.event.body.usage?.unit ?? internalModel?.unit)
-            : internalModel?.unit,
-        level: this.event.body.level ?? undefined,
-        statusMessage: this.event.body.statusMessage ?? undefined,
-        parentObservationId: this.event.body.parentObservationId ?? undefined,
-        version: this.event.body.version ?? undefined,
-        projectId: apiScope.projectId,
-        promptId: prompt ? prompt.id : undefined,
-        ...(internalModel
-          ? { internalModel: internalModel.modelName }
-          : undefined),
-        inputCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.inputCost
-            : undefined,
-        outputCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.outputCost
-            : undefined,
-        totalCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.totalCost
-            : undefined,
-        calculatedInputCost: calculatedCosts?.inputCost,
-        calculatedOutputCost: calculatedCosts?.outputCost,
-        calculatedTotalCost: calculatedCosts?.totalCost,
-        internalModelId: internalModel?.id,
-      },
-      update: {
-        name: this.event.body.name ?? undefined,
-        startTime: this.event.body.startTime
-          ? new Date(this.event.body.startTime)
-          : undefined,
-        endTime:
-          "endTime" in this.event.body && this.event.body.endTime
-            ? new Date(this.event.body.endTime)
-            : undefined,
-        completionStartTime:
-          "completionStartTime" in this.event.body &&
-          this.event.body.completionStartTime
-            ? new Date(this.event.body.completionStartTime)
-            : undefined,
-        metadata: mergedMetadata ?? this.event.body.metadata ?? undefined,
-        model: "model" in this.event.body ? this.event.body.model : undefined,
-        modelParameters:
-          "modelParameters" in this.event.body
-            ? (this.event.body.modelParameters ?? undefined)
-            : undefined,
-        input: this.event.body.input ?? undefined,
-        output: this.event.body.output ?? undefined,
-        promptTokens: newInputCount,
-        completionTokens: newOutputCount,
-        totalTokens: newTotalCount,
-        unit:
-          "usage" in this.event.body
-            ? (this.event.body.usage?.unit ?? internalModel?.unit)
-            : internalModel?.unit,
-        level: this.event.body.level ?? undefined,
-        statusMessage: this.event.body.statusMessage ?? undefined,
-        parentObservationId: this.event.body.parentObservationId ?? undefined,
-        version: this.event.body.version ?? undefined,
-        promptId: prompt ? prompt.id : undefined,
-        ...(internalModel
-          ? { internalModel: internalModel.modelName }
-          : undefined),
-        inputCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.inputCost
-            : undefined,
-        outputCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.outputCost
-            : undefined,
-        totalCost:
-          "usage" in this.event.body
-            ? this.event.body.usage?.totalCost
-            : undefined,
-        calculatedInputCost: calculatedCosts?.inputCost,
-        calculatedOutputCost: calculatedCosts?.outputCost,
-        calculatedTotalCost: calculatedCosts?.totalCost,
-        internalModelId: internalModel?.id,
-      },
-    };
-  }
-
-  async calculateTokenCounts(
-    projectId: string,
-    body:
-      | z.infer<typeof legacyObservationCreateEvent>["body"]
-      | z.infer<typeof generationCreateEvent>["body"],
-    calculateTokenDelegate: (p: {
-      model: Model;
-      text: unknown;
-    }) => number | undefined,
-    model?: Model,
-    existingObservation?: Omit<Observation, "input" | "output">,
-  ) {
-    let newPromptTokens = body.usage?.input || body.usageDetails?.input;
-    if (newPromptTokens === undefined && model && model.tokenizerId) {
-      if (body.input) {
-        newPromptTokens = calculateTokenDelegate({
-          model: model,
-          text: body.input,
-        });
-      } else {
-        logger.debug(
-          `No input provided, trying to calculate for id: ${existingObservation?.id}`,
-        );
-        const observationInput = await prisma.observation.findFirst({
-          where: { id: existingObservation?.id, projectId: projectId },
-          select: {
-            input: true,
-          },
-        });
-
-        newPromptTokens = calculateTokenDelegate({
-          model: model,
-          text: observationInput?.input,
-        });
-      }
-    }
-
-    let newCompletionTokens = body.usage?.output || body.usageDetails?.output;
-
-    if (newCompletionTokens === undefined && model && model.tokenizerId) {
-      if (body.output) {
-        newCompletionTokens = calculateTokenDelegate({
-          model: model,
-          text: body.output,
-        });
-      } else {
-        logger.debug(
-          `No output provided, trying to calculate for id: ${existingObservation?.id}`,
-        );
-        const observationOutput = await prisma.observation.findFirst({
-          where: { id: existingObservation?.id, projectId: projectId },
-          select: {
-            output: true,
-          },
-        });
-        newCompletionTokens = calculateTokenDelegate({
-          model: model,
-          text: observationOutput?.output,
-        });
-      }
-    }
-
-    return [newPromptTokens ?? undefined, newCompletionTokens ?? undefined];
-  }
-
-  static calculateTokenCosts(
-    model: Model | null | undefined,
-    userProvidedCosts: {
-      inputCost?: Decimal | null;
-      outputCost?: Decimal | null;
-      totalCost?: Decimal | null;
-    },
-    tokenCounts: { input?: number; output?: number; total?: number },
-  ): {
-    inputCost?: Decimal | null;
-    outputCost?: Decimal | null;
-    totalCost?: Decimal | null;
-  } {
-    // If user has provided any cost point, do not calculate anything else
-    if (
-      userProvidedCosts.inputCost ||
-      userProvidedCosts.outputCost ||
-      userProvidedCosts.totalCost
-    ) {
-      return {
-        ...userProvidedCosts,
-        totalCost:
-          userProvidedCosts.totalCost ??
-          (userProvidedCosts.inputCost ?? new Decimal(0)).add(
-            userProvidedCosts.outputCost ?? new Decimal(0),
-          ),
-      };
-    }
-
-    const finalInputCost =
-      tokenCounts.input !== undefined && model?.inputPrice
-        ? model.inputPrice.mul(tokenCounts.input)
-        : undefined;
-
-    const finalOutputCost =
-      tokenCounts.output !== undefined && model?.outputPrice
-        ? model.outputPrice.mul(tokenCounts.output)
-        : finalInputCost
-          ? new Decimal(0)
-          : undefined;
-
-    const finalTotalCost =
-      tokenCounts.total !== undefined && model?.totalPrice
-        ? model.totalPrice.mul(tokenCounts.total)
-        : (finalInputCost ?? finalOutputCost)
-          ? new Decimal(finalInputCost ?? 0).add(finalOutputCost ?? 0)
-          : undefined;
-
-    return {
-      inputCost: finalInputCost,
-      outputCost: finalOutputCost,
-      totalCost: finalTotalCost,
-    };
-  }
-
-  auth(apiScope: LegacyIngestionAccessScope): void {
-    if (apiScope.accessLevel !== "all")
-      throw new ForbiddenError("Access denied for observation creation");
-  }
-
-  async process(apiScope: LegacyIngestionAccessScope): Promise<Observation> {
-    this.auth(apiScope);
-
-    const existingObservation = this.event.body.id
-      ? await prisma.observation.findFirst({
-          select: {
-            // do not select I/O to spare our db
-            input: false,
-            output: false,
-
-            id: true,
-            traceId: true,
-            projectId: true,
-            type: true,
-            startTime: true,
-            endTime: true,
-            name: true,
-            metadata: true,
-            parentObservationId: true,
-            level: true,
-            statusMessage: true,
-            version: true,
-            createdAt: true,
-            updatedAt: true,
-            model: true,
-            internalModelId: true,
-            modelParameters: true,
-            promptTokens: true,
-            completionTokens: true,
-            totalTokens: true,
-            unit: true,
-            inputCost: true,
-            outputCost: true,
-            totalCost: true,
-            calculatedInputCost: true,
-            calculatedOutputCost: true,
-            calculatedTotalCost: true,
-            completionStartTime: true,
-            promptId: true,
-            internalModel: true,
-          },
-          where: { id: this.event.body.id, projectId: apiScope.projectId },
-        })
-      : null;
-
-    if (
-      existingObservation &&
-      existingObservation.projectId !== apiScope.projectId
-    ) {
-      throw new ForbiddenError(
-        `Access denied for observation creation ${existingObservation.projectId} `,
-      );
-    }
-
-    const obs = await this.convertToObservation(apiScope, existingObservation);
-
-    // Do not use nested upserts or multiple where conditions as this should be a single native database upsert
-    // https://www.prisma.io/docs/orm/reference/prisma-client-reference#database-upserts
-    return await prisma.observation.upsert({
-      where: {
-        id: obs.id,
-      },
-      create: obs.create,
-      update: obs.update,
-    });
-  }
-}
-
-export class TraceProcessor implements EventProcessor {
-  event: z.infer<typeof traceEvent>;
-
-  constructor(event: z.infer<typeof traceEvent>) {
-    this.event = event;
-  }
-
-  auth(apiScope: LegacyIngestionAccessScope): void {
-    if (apiScope.accessLevel !== "all")
-      throw new ForbiddenError("Access denied for trace creation");
-  }
-
-  async process(
-    apiScope: LegacyIngestionAccessScope,
-  ): Promise<Trace | Observation | Score> {
-    const { body } = this.event;
-
-    this.auth(apiScope);
-
-    const internalId =
-      body.id ??
-      (() => {
-        const newId = v4();
-        logger.info(
-          `trace.id is null. Generating for projectId: ${apiScope.projectId}, id: ${newId}`,
-        );
-        return newId;
-      })();
-
-    logger.debug(
-      `Trying to create trace, project ${apiScope.projectId}, id: ${internalId}`,
-    );
-
-    const existingTrace = await prisma.trace.findFirst({
-      where: {
-        id: internalId,
-      },
-    });
-
-    if (existingTrace && existingTrace.projectId !== apiScope.projectId) {
-      throw new ForbiddenError(
-        `Access denied for trace creation ${existingTrace.projectId}`,
-      );
-    }
-
-    const mergedMetadata = mergeJson(
-      existingTrace?.metadata
-        ? jsonSchema.parse(existingTrace.metadata)
-        : undefined,
-      body.metadata ?? undefined,
-    );
-
-    const mergedTags =
-      existingTrace?.tags && body.tags
-        ? Array.from(new Set(existingTrace.tags.concat(body.tags ?? []))).sort()
-        : body.tags
-          ? Array.from(new Set(body.tags)).sort()
-          : undefined;
-
-    if (body.sessionId) {
-      try {
-        await prisma.traceSession.upsert({
-          where: {
-            id_projectId: {
-              id: body.sessionId,
-              projectId: apiScope.projectId,
-            },
-          },
-          create: {
-            id: body.sessionId,
-            projectId: apiScope.projectId,
-          },
-          update: {},
-        });
-      } catch (e) {
-        if (
-          e instanceof Prisma.PrismaClientKnownRequestError &&
-          e.code === "P2002"
-        ) {
-          logger.warn(
-            `Failed to upsert session. Session ${body.sessionId} in project ${apiScope.projectId} already exists`,
-          );
-        } else {
-          throw e;
-        }
-      }
-    }
-
-    // Do not use nested upserts or multiple where conditions as this should be a single native database upsert
-    // https://www.prisma.io/docs/orm/reference/prisma-client-reference#database-upserts
-    const upsertedTrace = await prisma.trace.upsert({
-      where: {
-        id: internalId,
-      },
-      create: {
-        id: internalId,
-        timestamp: this.event.body.timestamp
-          ? new Date(this.event.body.timestamp)
-          : undefined,
-        name: body.name ?? undefined,
-        userId: body.userId ?? undefined,
-        input: body.input ?? undefined,
-        output: body.output ?? undefined,
-        metadata: mergedMetadata ?? body.metadata ?? undefined,
-        release: body.release ?? undefined,
-        version: body.version ?? undefined,
-        sessionId: body.sessionId ?? undefined,
-        public: body.public ?? undefined,
-        projectId: apiScope.projectId,
-        tags: mergedTags ?? undefined,
-      },
-      update: {
-        name: body.name ?? undefined,
-        timestamp: this.event.body.timestamp
-          ? new Date(this.event.body.timestamp)
-          : undefined,
-        userId: body.userId ?? undefined,
-        input: body.input ?? undefined,
-        output: body.output ?? undefined,
-        metadata: mergedMetadata ?? body.metadata ?? undefined,
-        release: body.release ?? undefined,
-        version: body.version ?? undefined,
-        sessionId: body.sessionId ?? undefined,
-        public: body.public ?? undefined,
-        tags: mergedTags ?? undefined,
-      },
-    });
-    return upsertedTrace;
-  }
-}
-
-export class ScoreProcessor implements EventProcessor {
-  event: z.infer<typeof scoreEvent>;
-
-  constructor(event: z.infer<typeof scoreEvent>) {
-    this.event = event;
-  }
-
-  auth(apiScope: LegacyIngestionAccessScope) {
-    if (apiScope.accessLevel !== "scores" && apiScope.accessLevel !== "all")
-      throw new ForbiddenError(
-        `Access denied for score creation, ${apiScope.accessLevel}`,
-      );
-  }
-
-  async process(
-    apiScope: LegacyIngestionAccessScope,
-  ): Promise<Trace | Observation | Score> {
-    const { body } = this.event;
-
-    this.auth(apiScope);
-
-    const id =
-      body.id ??
-      (() => {
-        const newId = v4();
-        logger.info(
-          `score.id is null. Generating for projectId: ${apiScope.projectId}, id: ${newId}`,
-        );
-        return newId;
-      })();
-
-    const existingScore = await prisma.score.findFirst({
-      where: {
-        id: id,
-      },
-      select: {
-        projectId: true,
-      },
-    });
-    if (existingScore && existingScore.projectId !== apiScope.projectId) {
-      throw new ForbiddenError(
-        `Access denied for score creation ${existingScore.projectId}`,
-      );
-    }
-
-    const validatedScore = await validateAndInflateScore({
-      body,
-      scoreId: id,
-      projectId: apiScope.projectId,
-    });
-
-    return await prisma.score.upsert({
-      where: {
-        id_projectId: {
-          id,
-          projectId: apiScope.projectId,
-        },
-      },
-      create: {
-        ...validatedScore,
-      },
-      update: {
-        ...validatedScore,
-      },
-    });
-  }
-}
-
-export class SdkLogProcessor implements EventProcessor {
-  event: z.infer<typeof sdkLogEvent>;
-
-  constructor(event: z.infer<typeof sdkLogEvent>) {
-    this.event = event;
-  }
-
-  auth(apiScope: LegacyIngestionAccessScope) {
-    return;
-  }
-
-  process() {
-    try {
-      logger.info("SDK Log", this.event);
-      return undefined;
-    } catch (error) {
-      return undefined;
-    }
-  }
-}

packages/shared/src/server/ingestion/legacy/index.ts

@@ -1,161 +0,0 @@
-import { env } from "node:process";
-import z from "zod";
-import { ForbiddenError, UnauthorizedError } from "../../../errors";
-import { eventTypes, ingestionApiSchema, IngestionEventType } from "../types";
-import { getProcessorForEvent } from "./EventProcessor";
-import { ApiAccessScope } from "../../auth/types";
-import { backOff } from "exponential-backoff";
-import { Model } from "../../..";
-import { logger } from "../../logger";
-
-export type BatchResult = {
-  result: unknown;
-  id: string;
-  type: string;
-};
-
-type TokenCountInput = {
-  model: Model;
-  text: unknown;
-};
-
-export type LegacyIngestionAccessScope = Omit<
-  ApiAccessScope,
-  "orgId" | "plan" | "rateLimitOverrides"
->;
-
-type LegacyIngestionAuthHeaderVerificationResult =
-  | {
-      validKey: true;
-      scope: LegacyIngestionAccessScope;
-    }
-  | {
-      validKey: false;
-      error: string;
-    };
-
-export const handleBatch = async (
-  events: z.infer<typeof ingestionApiSchema>["batch"],
-  authCheck: LegacyIngestionAuthHeaderVerificationResult,
-  calculateTokenDelegate: (p: TokenCountInput) => number | undefined,
-) => {
-  logger.debug(`handling ingestion ${events.length} events`);
-
-  if (!authCheck.validKey) throw new UnauthorizedError(authCheck.error);
-
-  const results: BatchResult[] = []; // Array to store the results
-
-  const errors: {
-    error: unknown;
-    id: string;
-    type: string;
-  }[] = []; // Array to store the errors
-
-  for (const singleEvent of events) {
-    try {
-      const result = await retry(async () => {
-        return await handleSingleEvent(
-          singleEvent,
-          authCheck.scope,
-          calculateTokenDelegate,
-        );
-      });
-      results.push({
-        result: result,
-        id: singleEvent.id,
-        type: singleEvent.type,
-      }); // Push each result into the array
-    } catch (error) {
-      // Handle or log the error if `handleSingleEvent` fails
-      logger.error("Error handling event:", error);
-      // Decide how to handle the error: rethrow, continue, or push an error object to results
-      // For example, push an error object:
-      errors.push({
-        error,
-        id: singleEvent.id,
-        type: singleEvent.type,
-      });
-    }
-  }
-
-  return { results, errors };
-};
-
-async function retry<T>(request: () => Promise<T>): Promise<T> {
-  return await backOff(request, {
-    numOfAttempts: 5,
-    retry: (e: Error, attemptNumber: number) => {
-      if (e instanceof UnauthorizedError || e instanceof ForbiddenError) {
-        logger.info("not retrying auth error");
-        return false;
-      }
-      logger.info(`retrying processing events ${attemptNumber}`);
-      return true;
-    },
-  });
-}
-
-const handleSingleEvent = async (
-  event: IngestionEventType,
-  apiScope: LegacyIngestionAccessScope,
-  calculateTokenDelegate: (p: {
-    model: Model;
-    text: unknown;
-  }) => number | undefined,
-) => {
-  const { body } = event;
-  let restEvent = body;
-  if ("input" in body) {
-    // eslint-disable-next-line @typescript-eslint/no-unused-vars
-    const { input, ...rest } = body;
-    restEvent = rest;
-  }
-  if ("output" in restEvent) {
-    // eslint-disable-next-line @typescript-eslint/no-unused-vars
-    const { output, ...rest } = restEvent;
-    restEvent = rest;
-  }
-
-  logger.debug(
-    `handling single event ${event.id} of type ${event.type}:  ${JSON.stringify({ body: restEvent })}`,
-  );
-
-  const cleanedEvent = cleanEvent(event) as IngestionEventType;
-
-  // Deny access to non-score events if the access level is not "all"
-  // This is an additional safeguard to auth checks in EventProcessor
-  if (
-    apiScope.accessLevel !== "all" &&
-    cleanedEvent.type !== eventTypes.SCORE_CREATE
-  ) {
-    throw new ForbiddenError("Access denied. Event type not allowed.");
-  }
-
-  return getProcessorForEvent(cleanedEvent, calculateTokenDelegate).process(
-    apiScope,
-  );
-};
-
-// cleans NULL characters from the event
-export function cleanEvent(obj: unknown): unknown {
-  if (typeof obj === "string") {
-    return obj.replace(/\u0000/g, "");
-  } else if (typeof obj === "object" && obj !== null) {
-    if (Array.isArray(obj)) {
-      return obj.map(cleanEvent);
-    } else {
-      // Here we assert that obj is a Record<string, unknown>
-      const objAsRecord = obj as Record<string, unknown>;
-      const newObj: Record<string, unknown> = {};
-      for (const key in objAsRecord) {
-        newObj[key] = cleanEvent(objAsRecord[key]);
-      }
-      return newObj;
-    }
-  } else {
-    return obj;
-  }
-}
-
-export const isUndefinedOrNull = <T>(val?: T | null): val is undefined | null =>
-  val === undefined || val === null;

packages/shared/src/server/ingestion/processEventBatch.ts

@@ -18,22 +18,14 @@ import {
   traceException,
 } from "../instrumentation";
 import { logger } from "../logger";
-import { LegacyIngestionEventType, QueueJobs } from "../queues";
+import { QueueJobs } from "../queues";
 import { IngestionQueue } from "../redis/ingestionQueue";
-import { LegacyIngestionQueue } from "../redis/legacyIngestion";
 import { redis } from "../redis/redis";
-import { handleBatch } from "./legacy";
+import { eventTypes, ingestionEvent, IngestionEventType } from "./types";
 import {
   StorageService,
   StorageServiceFactory,
 } from "../services/StorageService";
-import { getProcessorForEvent } from "./legacy/EventProcessor";
-import { eventTypes, ingestionEvent, IngestionEventType } from "./types";
-
-export type TokenCountDelegate = (p: {
-  model: Model;
-  text: unknown;
-}) => number | undefined;
 
 let s3StorageServiceClient: StorageService;
 
@@ -51,10 +43,45 @@ const getS3StorageServiceClient = (bucketName: string): StorageService => {
   return s3StorageServiceClient;
 };
 
+export type TokenCountDelegate = (p: {
+  model: Model;
+  text: unknown;
+}) => number | undefined;
+
+/**
+ * Get the delay for the event based on the event type. Uses delay if set, 0 if current UTC timestamp is not between
+ * 23:45 and 00:15, and env.LANGFUSE_INGESTION_QUEUE_DELAY_MS otherwise.
+ * We need the delay around date boundaries to avoid duplicates for out-of-order processing of events.
+ * @param delay - Delay overwrite. Used if non-null.
+ */
+const getDelay = (delay: number | null) => {
+  if (delay !== null) {
+    return delay;
+  }
+  const now = new Date();
+  const hours = now.getUTCHours();
+  const minutes = now.getUTCMinutes();
+
+  if ((hours === 23 && minutes >= 45) || (hours === 0 && minutes <= 15)) {
+    return env.LANGFUSE_INGESTION_QUEUE_DELAY_MS;
+  }
+
+  // Use 5s here to avoid duplicate processing on the worker. If the ingestion delay is set to a lower value,
+  // we use this instead.
+  // Values should be revisited based on a cost/performance trade-off.
+  return Math.min(5000, env.LANGFUSE_INGESTION_QUEUE_DELAY_MS);
+};
+
+/**
+ * Processes a batch of events.
+ * @param input - Batch of IngestionEventType. Will validate the types first thing and return errors if they are invalid.
+ * @param authCheck - AuthHeaderValidVerificationResult
+ * @param delay - (Optional) Delay in ms to wait before processing events in the batch.
+ */
 export const processEventBatch = async (
   input: unknown[],
   authCheck: AuthHeaderValidVerificationResult,
-  tokenCountDelegate: TokenCountDelegate,
+  delay: number | null = null,
 ): Promise<{
   successes: { id: string; status: number }[];
   errors: {
@@ -67,7 +94,10 @@ export const processEventBatch = async (
   // add context of api call to the span
   const currentSpan = getCurrentSpan();
   recordIncrement("langfuse.ingestion.event", input.length);
-  currentSpan?.setAttribute("event_count", input.length);
+  currentSpan?.setAttribute("langfuse.ingestion.batch_size", input.length);
+  currentSpan?.setAttribute("langfuse.project.id", authCheck.scope.projectId);
+  currentSpan?.setAttribute("langfuse.org.id", authCheck.scope.orgId);
+  currentSpan?.setAttribute("langfuse.org.plan", authCheck.scope.plan);
 
   /**************
    * VALIDATION *
@@ -82,7 +112,10 @@ export const processEventBatch = async (
         (span) => {
           const parsedBody = ingestionEvent.safeParse(event);
           if (parsedBody.data?.id !== undefined) {
-            span.setAttribute("object.id", parsedBody.data.id);
+            span.setAttribute(
+              "langfuse.ingestion.entity.id",
+              parsedBody.data.id,
+            );
           }
           return parsedBody;
         },
@@ -99,7 +132,7 @@ export const processEventBatch = async (
         });
         return [];
       }
-      if (!isAuthorized(parsed.data, authCheck, tokenCountDelegate)) {
+      if (!isAuthorized(parsed.data, authCheck)) {
         authenticationErrors.push({
           id: parsed.data.id,
           error: new UnauthorizedError("Access Scope Denied"),
@@ -157,9 +190,6 @@ export const processEventBatch = async (
    ********************/
   let s3UploadErrored = false;
   await instrumentAsync({ name: "s3-upload-events" }, async () => {
-    const s3Client = getS3StorageServiceClient(
-      env.LANGFUSE_S3_EVENT_UPLOAD_BUCKET,
-    );
     // S3 Event Upload is blocking, but non-failing.
     // If a promise rejects, we log it below, but do not throw an error.
     // In this case, we upload the full batch into the Redis queue.
@@ -169,10 +199,10 @@ export const processEventBatch = async (
         // That way we batch updates from the same invocation into a single file and reduce
         // write operations on S3.
         const { data, key, type, eventBodyId } = sortedBatchByEventBodyId[id];
-        return s3Client.uploadJson(
-          `${env.LANGFUSE_S3_EVENT_UPLOAD_PREFIX}${authCheck.scope.projectId}/${getClickhouseEntityType(type)}/${eventBodyId}/${key}.json`,
-          data,
-        );
+        const bucketPath = `${env.LANGFUSE_S3_EVENT_UPLOAD_PREFIX}${authCheck.scope.projectId}/${getClickhouseEntityType(type)}/${eventBodyId}/${key}.json`;
+        return getS3StorageServiceClient(
+          env.LANGFUSE_S3_EVENT_UPLOAD_BUCKET,
+        ).uploadJson(bucketPath, data);
       }),
     );
     results.forEach((result) => {
@@ -185,154 +215,65 @@ export const processEventBatch = async (
     });
   });
 
-  // This is a workaround to allow us to disable async ingestion processing for SDK CI testing
-  // TODO: remove this block after SDKs are ready for V3 async ingestion processing
-  if (env.LANGFUSE_SDK_CI_SYNC_PROCESSING_ENABLED === "true") {
-    const result = await handleBatch(
-      sortedBatch,
-      authCheck,
-      tokenCountDelegate,
-    );
-
-    //  in case we did not return early, we return the result here
-    return aggregateBatchResult(
-      [...validationErrors, ...authenticationErrors, ...result.errors],
-      result.results,
-      authCheck.scope.projectId,
-    );
-  }
-
   // Send each event individually to IngestionQueue for ClickHouse processing
-  if (env.LANGFUSE_CLICKHOUSE_INGESTION_ENABLED === "true") {
-    if (s3UploadErrored) {
-      throw new Error(
-        "Failed to upload events to blob storage, aborting event processing",
-      );
-    }
-
-    if (redis) {
-      const queue = IngestionQueue.getInstance();
-      await Promise.all(
-        Object.keys(sortedBatchByEventBodyId).map(async (id) =>
-          queue
-            ? queue.add(
-                QueueJobs.IngestionJob,
-                {
-                  id: randomUUID(),
-                  timestamp: new Date(),
-                  name: QueueJobs.IngestionJob as const,
-                  payload: {
-                    data: {
-                      type: sortedBatchByEventBodyId[id].type,
-                      eventBodyId: sortedBatchByEventBodyId[id].eventBodyId,
-                    },
-                    authCheck,
-                  },
-                },
-                {
-                  delay: env.LANGFUSE_INGESTION_QUEUE_DELAY_MS,
-                },
-              )
-            : Promise.reject("Failed to instantiate queue"),
-        ),
-      );
-      if (env.LANGFUSE_POSTGRES_INGESTION_ENABLED !== "true") {
-        // If postgres ingestion is disabled, we return early
-        return aggregateBatchResult(
-          [...validationErrors, ...authenticationErrors],
-          sortedBatch.map((event) => ({ id: event.id, result: event })),
-          authCheck.scope.projectId,
-        );
-      }
-    }
-  }
-
-  if (env.LANGFUSE_POSTGRES_INGESTION_ENABLED === "true") {
-    // As part of the legacy processing we sent the entire batch to the worker.
-    if (redis) {
-      const queue = LegacyIngestionQueue.getInstance();
-
-      if (queue) {
-        let addToQueueFailed = false;
-
-        const queuePayload: LegacyIngestionEventType = !s3UploadErrored
-          ? {
-              data: Object.keys(sortedBatchByEventBodyId).map((id) => {
-                const { key, type, eventBodyId } = sortedBatchByEventBodyId[id];
-                return {
-                  type,
-                  eventBodyId,
-                  eventId: key,
-                };
-              }),
-              authCheck,
-              useS3EventStore: true,
-            }
-          : { data: sortedBatch, authCheck, useS3EventStore: false };
-
-        try {
-          await queue.add(QueueJobs.LegacyIngestionJob, {
-            payload: queuePayload,
-            id: randomUUID(),
-            timestamp: new Date(),
-            name: QueueJobs.LegacyIngestionJob as const,
-          });
-        } catch (e: unknown) {
-          logger.warn(
-            "Failed to add batch to queue, falling back to sync processing",
-            e,
-          );
-          addToQueueFailed = true;
-        }
-
-        if (!addToQueueFailed) {
-          return aggregateBatchResult(
-            // we are not sending additional server errors to the client in case of early return
-            [...validationErrors, ...authenticationErrors],
-            sortedBatch.map((event) => ({ id: event.id, result: event })),
-            authCheck.scope.projectId,
-          );
-        }
-      } else {
-        logger.warn(
-          "Ingestion queue not initialized, falling back to sync processing",
-        );
-      }
-    }
-
-    /*******************
-     * SYNC PROCESSING *
-     *******************/
-    const result = await handleBatch(
-      sortedBatch,
-      authCheck,
-      tokenCountDelegate,
-    );
-
-    //  in case we did not return early, we return the result here
-    return aggregateBatchResult(
-      [...validationErrors, ...authenticationErrors, ...result.errors],
-      result.results,
-      authCheck.scope.projectId,
+  if (s3UploadErrored) {
+    throw new Error(
+      "Failed to upload events to blob storage, aborting event processing",
     );
   }
 
-  throw new Error(
-    "Either Clickhouse or Postgres ingestion (or both) must be enabled",
+  if (!redis) {
+    throw new Error("Redis not initialized, aborting event processing");
+  }
+
+  const queue = IngestionQueue.getInstance();
+  await Promise.all(
+    Object.keys(sortedBatchByEventBodyId).map(async (id) =>
+      queue
+        ? queue.add(
+            QueueJobs.IngestionJob,
+            {
+              id: randomUUID(),
+              timestamp: new Date(),
+              name: QueueJobs.IngestionJob as const,
+              payload: {
+                data: {
+                  type: sortedBatchByEventBodyId[id].type,
+                  eventBodyId: sortedBatchByEventBodyId[id].eventBodyId,
+                  fileKey: sortedBatchByEventBodyId[id].key,
+                },
+                authCheck,
+              },
+            },
+            { delay: getDelay(delay) },
+          )
+        : Promise.reject("Failed to instantiate queue"),
+    ),
+  );
+
+  return aggregateBatchResult(
+    [...validationErrors, ...authenticationErrors],
+    sortedBatch.map((event) => ({ id: event.id, result: event })),
+    authCheck.scope.projectId,
   );
 };
 
 const isAuthorized = (
   event: IngestionEventType,
   authScope: AuthHeaderValidVerificationResult,
-  tokenCountDelegate: TokenCountDelegate,
 ): boolean => {
-  try {
-    getProcessorForEvent(event, tokenCountDelegate).auth(authScope.scope);
+  if (event.type === eventTypes.SDK_LOG) {
     return true;
-  } catch (error) {
-    return false;
   }
+
+  if (event.type === eventTypes.SCORE_CREATE) {
+    return (
+      authScope.scope.accessLevel === "scores" ||
+      authScope.scope.accessLevel === "all"
+    );
+  }
+
+  return authScope.scope.accessLevel === "all";
 };
 
 /**

packages/shared/src/server/ingestion/types.ts

@@ -66,8 +66,12 @@ const OpenAIUsageSchema = z
     prompt_tokens: z.number().nonnegative(),
     completion_tokens: z.number().nonnegative(),
     total_tokens: z.number().nonnegative(),
-    prompt_tokens_details: z.record(z.string(), z.number().nonnegative()),
-    completion_tokens_details: z.record(z.string(), z.number().nonnegative()),
+    prompt_tokens_details: z
+      .record(z.string(), z.number().nonnegative())
+      .nullish(),
+    completion_tokens_details: z
+      .record(z.string(), z.number().nonnegative())
+      .nullish(),
   })
   .strict()
   .transform((v) => {
@@ -112,13 +116,16 @@ export const UsageOrCostDetails = z
   .union([OpenAIUsageSchema, RawUsageOrCostDetails])
   .nullish();
 
+// Using z.any instead of jsonSchema for input/output as we saw huge CPU overhead for large numeric arrays.
+// With this setup parsing should be more lightweight and doesn't block other requests.
+// As we allow plain values, arrays, and objects the JSON parse via bodyParser should suffice.
 export const TraceBody = z.object({
   id: z.string().nullish(),
   timestamp: stringDateTime,
   name: z.string().max(1000).nullish(),
   externalId: z.string().nullish(),
-  input: jsonSchema.nullish(),
-  output: jsonSchema.nullish(),
+  input: z.any().nullish(),
+  output: z.any().nullish(),
   sessionId: z.string().nullish(),
   userId: z.string().nullish(),
   metadata: jsonSchema.nullish(),
@@ -133,8 +140,8 @@ export const OptionalObservationBody = z.object({
   name: z.string().nullish(),
   startTime: stringDateTime,
   metadata: jsonSchema.nullish(),
-  input: jsonSchema.nullish(),
-  output: jsonSchema.nullish(),
+  input: z.any().nullish(),
+  output: z.any().nullish(),
   level: z.nativeEnum(ObservationLevel).nullish(),
   statusMessage: z.string().nullish(),
   parentObservationId: z.string().nullish(),

packages/shared/src/server/instrumentation/index.ts

@@ -226,6 +226,14 @@ export const recordHistogram = (
   dd.dogstatsd.histogram(stat, value, tags);
 };
 
+export const recordDistribution = (
+  stat: string,
+  value?: number | undefined,
+  tags?: { [tag: string]: string | number } | undefined,
+) => {
+  dd.dogstatsd.distribution(stat, value, tags);
+};
+
 /**
  * Converts a queue name to the matching datadog metric name.
  * Consumer only needs to append the relevant suffix.

packages/shared/src/server/llm/fetchLLMCompletion.ts

@@ -1,7 +1,5 @@
 import type { ZodSchema } from "zod";
 
-import { CallbackHandler } from "langfuse-langchain";
-
 import { ChatAnthropic } from "@langchain/anthropic";
 import { ChatVertexAI } from "@langchain/google-vertexai";
 import { ChatBedrockConverse } from "@langchain/aws";
@@ -21,27 +19,20 @@ import GCPServiceAccountKeySchema, {
   BedrockConfigSchema,
   BedrockCredentialSchema,
 } from "../../interfaces/customLLMProviderConfigSchemas";
-import { AuthHeaderValidVerificationResult } from "../auth/types";
-import {
-  processEventBatch,
-  type TokenCountDelegate,
-} from "../ingestion/processEventBatch";
+import { processEventBatch } from "../ingestion/processEventBatch";
 import { logger } from "../logger";
-import { ChatMessage, ChatMessageRole, LLMAdapter, ModelParams } from "./types";
-
+import {
+  ChatMessage,
+  ChatMessageRole,
+  LLMAdapter,
+  ModelParams,
+  TraceParams,
+} from "./types";
+import { CallbackHandler } from "langfuse-langchain";
 import type { BaseCallbackHandler } from "@langchain/core/callbacks/base";
 
 type ProcessTracedEvents = () => Promise<void>;
 
-export type TraceParams = {
-  traceName: string;
-  traceId: string;
-  projectId: string;
-  tags: string[];
-  tokenCountDelegate: TokenCountDelegate;
-  authCheck: AuthHeaderValidVerificationResult;
-};
-
 type LLMCompletionParams = {
   messages: ChatMessage[];
   modelParams: ModelParams;
@@ -49,9 +40,11 @@ type LLMCompletionParams = {
   callbacks?: BaseCallbackHandler[];
   baseURL?: string;
   apiKey: string;
+  extraHeaders?: Record<string, string>;
   maxRetries?: number;
   config?: Record<string, string> | null;
   traceParams?: TraceParams;
+  throwOnError?: boolean; // default is true
 };
 
 type FetchLLMCompletionParams = LLMCompletionParams & {
@@ -100,6 +93,8 @@ export async function fetchLLMCompletion(
     maxRetries,
     config,
     traceParams,
+    extraHeaders,
+    throwOnError = true,
   } = params;
 
   let finalCallbacks: BaseCallbackHandler[] | undefined = callbacks ?? [];
@@ -111,7 +106,6 @@ export async function fetchLLMCompletion(
       _isLocalEventExportEnabled: true,
       tags: traceParams.tags,
     });
-
     finalCallbacks.push(handler);
 
     processTracedEvents = async () => {
@@ -122,7 +116,6 @@ export async function fetchLLMCompletion(
         await processEventBatch(
           JSON.parse(JSON.stringify(events)), // stringify to emulate network event batch from network call
           traceParams.authCheck,
-          traceParams.tokenCountDelegate,
         );
       } catch (e) {
         logger.error("Failed to process traced events", { error: e });
@@ -177,6 +170,7 @@ export async function fetchLLMCompletion(
       maxRetries,
       configuration: {
         baseURL,
+        defaultHeaders: extraHeaders,
       },
       timeout: 1000 * 60 * 2, // 2 minutes timeout
     });
@@ -237,17 +231,18 @@ export async function fetchLLMCompletion(
     runName: traceParams?.traceName,
   };
 
-  if (params.structuredOutputSchema) {
-    return {
-      completion: await (chatModel as ChatOpenAI) // Typecast necessary due to https://github.com/langchain-ai/langchainjs/issues/6795
-        .withStructuredOutput(params.structuredOutputSchema)
-        .invoke(finalMessages, runConfig),
-      processTracedEvents,
-    };
-  }
+  try {
+    if (params.structuredOutputSchema) {
+      return {
+        completion: await (chatModel as ChatOpenAI) // Typecast necessary due to https://github.com/langchain-ai/langchainjs/issues/6795
+          .withStructuredOutput(params.structuredOutputSchema)
+          .invoke(finalMessages, runConfig),
+        processTracedEvents,
+      };
+    }
 
-  /*
-  Workaround OpenAI o1 while in beta:
+    /*
+  Workaround OpenAI reasoning models:
   
   This is a temporary workaround to avoid sending system messages to OpenAI's O1 models.
   O1 models do not support in beta:
@@ -258,43 +253,55 @@ export async function fetchLLMCompletion(
 
   Reference: https://platform.openai.com/docs/guides/reasoning/beta-limitations
   */
-  if (modelParams.model.startsWith("o1-")) {
-    return {
-      completion: await new ChatOpenAI({
-        openAIApiKey: apiKey,
-        modelName: modelParams.model,
-        temperature: 1,
-        maxTokens: undefined,
-        topP: undefined,
-        callbacks,
-        maxRetries,
-        configuration: {
-          baseURL,
-        },
-        timeout: 1000 * 60 * 2, // 2 minutes timeout
-      })
-        .pipe(new StringOutputParser())
-        .invoke(
-          finalMessages.filter((message) => message._getType() !== "system"),
-          runConfig,
-        ),
-      processTracedEvents,
-    };
-  }
+    if (
+      modelParams.model.startsWith("o1-") ||
+      modelParams.model.startsWith("o3-")
+    ) {
+      return {
+        completion: await new ChatOpenAI({
+          openAIApiKey: apiKey,
+          modelName: modelParams.model,
+          temperature: 1,
+          maxTokens: undefined,
+          topP: undefined,
+          callbacks,
+          maxRetries,
+          modelKwargs: {
+            max_completion_tokens: modelParams.max_tokens,
+          },
+          configuration: {
+            baseURL,
+          },
+          timeout: 1000 * 60 * 2, // 2 minutes timeout
+        })
+          .pipe(new StringOutputParser())
+          .invoke(
+            finalMessages.filter((message) => message._getType() !== "system"),
+            runConfig,
+          ),
+        processTracedEvents,
+      };
+    }
+
+    if (streaming) {
+      return {
+        completion: await chatModel
+          .pipe(new BytesOutputParser())
+          .stream(finalMessages, runConfig),
+        processTracedEvents,
+      };
+    }
 
-  if (streaming) {
     return {
       completion: await chatModel
-        .pipe(new BytesOutputParser())
-        .stream(finalMessages, runConfig),
+        .pipe(new StringOutputParser())
+        .invoke(finalMessages, runConfig),
       processTracedEvents,
     };
+  } catch (error) {
+    if (throwOnError) {
+      throw error;
+    }
+    return { completion: null, processTracedEvents };
   }
-
-  return {
-    completion: await chatModel
-      .pipe(new StringOutputParser())
-      .invoke(finalMessages, runConfig),
-    processTracedEvents,
-  };
 }

packages/shared/src/server/llm/types.ts

@@ -1,6 +1,8 @@
 import { LlmApiKeys } from "@prisma/client";
 import z from "zod";
 import { BedrockConfigSchema } from "../../interfaces/customLLMProviderConfigSchemas";
+import { TokenCountDelegate } from "../ingestion/processEventBatch";
+import { AuthHeaderValidVerificationResult } from "../auth/types";
 
 export type PromptVariable = { name: string; value: string; isUsed: boolean };
 
@@ -16,7 +18,7 @@ export enum LLMAdapter {
   OpenAI = "openai",
   Azure = "azure",
   Bedrock = "bedrock",
-  VertexAI = "vertex-ai",
+  VertexAI = "google-vertex-ai",
 }
 
 export enum ChatMessageRole {
@@ -71,6 +73,7 @@ export const ExperimentMetadataSchema = z
     provider: z.string(),
     model: z.string(),
     model_params: ZodModelConfig,
+    error: z.string().optional(),
   })
   .strict();
 export type ExperimentMetadata = z.infer<typeof ExperimentMetadataSchema>;
@@ -82,6 +85,8 @@ export const openAIModels = [
   "gpt-4o-2024-05-13",
   "gpt-4o-mini",
   "gpt-4o-mini-2024-07-18",
+  "o3-mini",
+  "o3-mini-2025-01-31",
   "o1-preview",
   "o1-preview-2024-09-12",
   "o1-mini",
@@ -116,6 +121,10 @@ export const anthropicModels = [
 ] as const;
 
 export const vertexAIModels = [
+  "gemini-2.0-pro-exp-02-05",
+  "gemini-2.0-flash-001",
+  "gemini-2.0-flash-lite-preview-02-05",
+  "gemini-2.0-flash-exp",
   "gemini-1.5-pro",
   "gemini-1.5-flash",
   "gemini-1.0-pro",
@@ -147,6 +156,8 @@ export const LLMApiKeySchema = z
     provider: z.string(),
     displaySecretKey: z.string(),
     secretKey: z.string(),
+    extraHeaders: z.string().nullish(),
+    extraHeaderKeys: z.array(z.string()),
     baseURL: z.string().nullable(),
     customModels: z.array(z.string()),
     withDefaultModels: z.boolean(),
@@ -160,3 +171,12 @@ export type LLMApiKey =
   z.infer<typeof LLMApiKeySchema> extends LlmApiKeys
     ? z.infer<typeof LLMApiKeySchema>
     : never;
+
+export type TraceParams = {
+  traceName: string;
+  traceId: string;
+  projectId: string;
+  tags: string[];
+  tokenCountDelegate: TokenCountDelegate;
+  authCheck: AuthHeaderValidVerificationResult;
+};

packages/shared/src/server/llm/utils.ts

@@ -0,0 +1,13 @@
+import { z } from "zod";
+
+import { decrypt } from "../../encryption";
+
+const ExtraHeaderSchema = z.record(z.string(), z.string());
+
+export function decryptAndParseExtraHeaders(
+  extraHeaders: string | null | undefined,
+) {
+  if (!extraHeaders) return;
+
+  return ExtraHeaderSchema.parse(JSON.parse(decrypt(extraHeaders)));
+}

packages/shared/src/server/queries/clickhouse-sql/clickhouse-filter.ts

@@ -383,6 +383,10 @@ export class FilterList {
     return this.filters.find(predicate);
   }
 
+  some(predicate: (filter: Filter) => boolean) {
+    return this.filters.some(predicate);
+  }
+
   length() {
     return this.filters.length;
   }

packages/shared/src/server/queries/createGenerationsQuery.ts

@@ -1,11 +1,4 @@
-import { ObservationView, Prisma } from "@prisma/client";
-import {
-  datetimeFilterToPrismaSql,
-  tableColumnsToSqlFilterAndPrefix,
-} from "../filterToPrisma";
-import { orderByToPrismaSql } from "../orderByToPrisma";
-import { observationsTableCols } from "../../observationsTable";
-import { TableFilters } from "./types";
+import { ObservationView } from "@prisma/client";
 
 type AdditionalObservationFields = {
   traceName: string | null;
@@ -26,147 +19,3 @@ export type IOAndMetadataOmittedObservations = Array<
   Omit<ObservationView, "input" | "output" | "metadata"> &
     AdditionalObservationFields
 >;
-
-export function parseGetAllGenerationsInput(filters: TableFilters) {
-  const searchCondition = filters.searchQuery
-    ? Prisma.sql`AND (
-        o."id" ILIKE ${`%${filters.searchQuery}%`} OR
-        o."name" ILIKE ${`%${filters.searchQuery}%`} OR
-        o."model" ILIKE ${`%${filters.searchQuery}%`} OR
-        t."name" ILIKE ${`%${filters.searchQuery}%`}
-      )`
-    : Prisma.empty;
-
-  const filterCondition = tableColumnsToSqlFilterAndPrefix(
-    filters.filter ?? [],
-    observationsTableCols,
-    "observations",
-  );
-
-  const orderByCondition = orderByToPrismaSql(
-    filters.orderBy,
-    observationsTableCols,
-  );
-
-  // to improve query performance, add timeseries filter to observation queries as well
-  const startTimeFilter = filters.filter?.find(
-    (f) => f.column === "Start Time" && f.type === "datetime",
-  );
-
-  const datetimeFilter =
-    startTimeFilter && startTimeFilter.type === "datetime"
-      ? datetimeFilterToPrismaSql(
-          "start_time",
-          startTimeFilter.operator,
-          startTimeFilter.value,
-        )
-      : Prisma.empty;
-
-  return {
-    searchCondition,
-    filterCondition,
-    orderByCondition,
-    datetimeFilter,
-  };
-}
-
-export function createGenerationsQuery({
-  projectId,
-  datetimeFilter = Prisma.empty,
-  page,
-  limit,
-  searchCondition = Prisma.empty,
-  filterCondition = Prisma.empty,
-  orderByCondition = Prisma.empty,
-  selectIOAndMetadata = false,
-  selectScoreValues = false,
-}: {
-  projectId: string;
-  datetimeFilter?: Prisma.Sql;
-  page?: number;
-  limit?: number;
-  searchCondition?: Prisma.Sql;
-  filterCondition?: Prisma.Sql;
-  orderByCondition?: Prisma.Sql;
-  selectIOAndMetadata?: boolean;
-  selectScoreValues?: boolean;
-}) {
-  return Prisma.sql`
-  WITH scores_avg AS (
-        SELECT
-          trace_id,
-          observation_id,
-          ${selectScoreValues ? Prisma.sql`jsonb_object_agg(name::text, "values") AS "scores_values",` : Prisma.empty}
-          jsonb_object_agg(name::text, avg_value::double precision) AS "scores_avg"
-        FROM (
-          SELECT
-            trace_id,
-            observation_id,
-            name,
-            ${selectScoreValues ? Prisma.sql`array_agg(COALESCE(string_value, value::text)) AS "values",` : Prisma.empty}
-            avg(value) avg_value,
-            comment
-          FROM
-            scores
-          WHERE
-            project_id = ${projectId}
-            ${selectScoreValues ? Prisma.empty : Prisma.sql`AND scores."data_type" IN ('NUMERIC', 'BOOLEAN')`}
-          GROUP BY
-            trace_id,
-            observation_id,
-            name,
-            comment
-          ORDER BY
-            trace_id
-          ) tmp
-        GROUP BY
-          trace_id, 
-          observation_id
-      )
-      SELECT
-        ${selectScoreValues ? Prisma.sql`s_avg."scores_values" AS "scores",` : Prisma.empty}
-        o.id,
-        o.name,
-        o.model,
-        o."modelParameters",
-        o.start_time as "startTime",
-        o.end_time as "endTime",
-        ${selectIOAndMetadata ? Prisma.sql`o.input, o.output, o.metadata,` : Prisma.empty} 
-        o.trace_id as "traceId",
-        t.name as "traceName",
-        o.completion_start_time as "completionStartTime",
-        o.time_to_first_token as "timeToFirstToken",
-        o.prompt_tokens as "promptTokens",
-        o.completion_tokens as "completionTokens",
-        o.total_tokens as "totalTokens",
-        o.unit,
-        o.level,
-        o.status_message as "statusMessage",
-        o.version,
-        o.model_id as "modelId",
-        o.input_price as "inputPrice",
-        o.output_price as "outputPrice",
-        o.total_price as "totalPrice",
-        o.calculated_input_cost as "calculatedInputCost",
-        o.calculated_output_cost as "calculatedOutputCost",
-        o.calculated_total_cost as "calculatedTotalCost",
-        o."latency",
-        o.prompt_id as "promptId",
-        p.name as "promptName",
-        p.version as "promptVersion",
-        t.tags as "traceTags"
-      FROM observations_view o
-      JOIN traces t ON t.id = o.trace_id AND t.project_id = ${projectId}
-      LEFT JOIN scores_avg AS s_avg ON s_avg.trace_id = t.id and s_avg.observation_id = o.id
-      LEFT JOIN prompts p ON p.id = o.prompt_id AND p.project_id = ${projectId}
-      WHERE
-        o.project_id = ${projectId}
-        AND o.type = 'GENERATION'
-        ${datetimeFilter}
-        ${searchCondition}
-        ${filterCondition}
-        ${orderByCondition}
-      ${limit ? Prisma.sql`LIMIT ${limit}` : Prisma.empty}
-      ${page && limit ? Prisma.sql`OFFSET ${page * limit}` : Prisma.empty}
-    `;
-}

packages/shared/src/server/queries/createTracesQuery.ts

@@ -1,135 +0,0 @@
-import { Prisma } from "@prisma/client";
-import { TableFilters } from "./types";
-import {
-  datetimeFilterToPrismaSql,
-  tableColumnsToSqlFilterAndPrefix,
-} from "../filterToPrisma";
-import { tracesTableCols } from "../../tableDefinitions/tracesTable";
-import { orderByToPrismaSql } from "../orderByToPrisma";
-
-export function parseTraceAllFilters(input: TableFilters) {
-  const filterCondition = tableColumnsToSqlFilterAndPrefix(
-    input.filter ?? [],
-    tracesTableCols,
-    "traces",
-  );
-  const orderByCondition = orderByToPrismaSql(input.orderBy, tracesTableCols);
-
-  // to improve query performance, add timeseries filter to observation queries as well
-  const timeseriesFilter = input.filter?.find(
-    (f) => f.column === "Timestamp" && f.type === "datetime",
-  );
-
-  const observationTimeseriesFilter =
-    timeseriesFilter && timeseriesFilter.type === "datetime"
-      ? datetimeFilterToPrismaSql(
-          "start_time",
-          timeseriesFilter.operator,
-          timeseriesFilter.value,
-        )
-      : Prisma.empty;
-
-  const searchCondition = input.searchQuery
-    ? Prisma.sql`AND (
-    t."id" ILIKE ${`%${input.searchQuery}%`} OR 
-    t."external_id" ILIKE ${`%${input.searchQuery}%`} OR 
-    t."user_id" ILIKE ${`%${input.searchQuery}%`} OR 
-    t."name" ILIKE ${`%${input.searchQuery}%`}
-  )`
-    : Prisma.empty;
-
-  return {
-    filterCondition,
-    orderByCondition,
-    observationTimeseriesFilter,
-    searchCondition,
-  };
-}
-
-export function createTracesQuery({
-  select,
-  projectId,
-  observationTimeseriesFilter = Prisma.empty,
-  page,
-  limit,
-  searchCondition = Prisma.empty,
-  filterCondition = Prisma.empty,
-  orderByCondition = Prisma.empty,
-  selectScoreValues = false,
-}: {
-  select: Prisma.Sql;
-  projectId: string;
-  observationTimeseriesFilter?: Prisma.Sql;
-  page?: number;
-  limit?: number;
-  searchCondition?: Prisma.Sql;
-  filterCondition?: Prisma.Sql;
-  orderByCondition?: Prisma.Sql;
-  selectScoreValues?: boolean;
-}) {
-  return Prisma.sql`
-  SELECT
-      ${select}
-  FROM
-    "traces" AS t
-  LEFT JOIN LATERAL (
-    SELECT
-      SUM(prompt_tokens) AS "promptTokens",
-      SUM(completion_tokens) AS "completionTokens",
-      SUM(total_tokens) AS "totalTokens",
-      SUM(calculated_total_cost) AS "calculatedTotalCost",
-      SUM(calculated_input_cost) AS "calculatedInputCost",
-      SUM(calculated_output_cost) AS "calculatedOutputCost"
-    FROM
-      "observations_view"
-    WHERE
-      trace_id = t.id
-      AND "type" = 'GENERATION'
-      AND "project_id" = ${projectId}
-      ${observationTimeseriesFilter}
-  ) AS generation_metrics ON true
-  LEFT JOIN LATERAL (
-    SELECT
-      COUNT(*) AS "observationCount",
-      EXTRACT(EPOCH FROM COALESCE(MAX("end_time"), MAX("start_time"))) - EXTRACT(EPOCH FROM MIN("start_time"))::double precision AS "latency",
-      COALESCE(  
-        MAX(CASE WHEN level = 'ERROR' THEN 'ERROR' END),  
-        MAX(CASE WHEN level = 'WARNING' THEN 'WARNING' END),  
-        MAX(CASE WHEN level = 'DEFAULT' THEN 'DEFAULT' END),  
-        'DEBUG'  
-      ) AS "level"
-    FROM
-        "observations"
-    WHERE
-        trace_id = t.id
-        AND "project_id" = ${projectId}
-         ${observationTimeseriesFilter}
-  ) AS observation_metrics ON true
-  LEFT JOIN LATERAL (
-    SELECT
-      ${selectScoreValues ? Prisma.sql`jsonb_object_agg(name::text, "values") AS "scores_values",` : Prisma.empty}
-      jsonb_object_agg(name::text, avg_value::double precision) AS "scores_avg"
-    FROM (
-        SELECT
-            name,
-            ${selectScoreValues ? Prisma.sql`array_agg(COALESCE(string_value, value::text)) AS "values",` : Prisma.empty}
-            AVG(value) avg_value
-        FROM
-            scores
-        WHERE
-            trace_id = t.id
-            AND t."project_id" = ${projectId}     
-            ${selectScoreValues ? Prisma.empty : Prisma.sql`AND scores."data_type" IN ('NUMERIC', 'BOOLEAN')`}
-        GROUP BY
-            name
-    ) tmp
-  ) AS s_avg ON true
-  WHERE 
-    t."project_id" = ${projectId}
-    ${searchCondition}
-    ${filterCondition}
-  ${orderByCondition}
-  ${limit ? Prisma.sql`LIMIT ${limit}` : Prisma.empty}
-  ${page !== undefined && limit !== undefined ? Prisma.sql`OFFSET ${page * limit}` : Prisma.empty}
-`;
-}

packages/shared/src/server/queries/index.ts

@@ -1,8 +1,5 @@
 export { createSessionsAllQuery } from "./createSessionsAllQuery";
-export { createTracesQuery, parseTraceAllFilters } from "./createTracesQuery";
 export {
-  createGenerationsQuery,
-  parseGetAllGenerationsInput,
   type FullObservations,
   type FullObservationsWithScores,
   type IOAndMetadataOmittedObservations,

packages/shared/src/server/queues.ts

@@ -1,45 +1,16 @@
 import { z } from "zod";
-import { eventTypes, ingestionBatchEvent } from ".";
-
-export const LegacyIngestionEventFull = z.object({
-  useS3EventStore: z.literal(false),
-  data: ingestionBatchEvent,
-  authCheck: z.object({
-    validKey: z.literal(true),
-    scope: z.object({
-      projectId: z.string(),
-      accessLevel: z.enum(["all", "scores"]),
-    }),
-  }),
-});
-
-export const LegacyIngestionEventMeta = z.object({
-  useS3EventStore: z.literal(true),
-  data: z.array(
-    z.object({
-      type: z.nativeEnum(eventTypes),
-      eventBodyId: z.string(),
-      eventId: z.string(),
-    }),
-  ),
-  authCheck: z.object({
-    validKey: z.literal(true),
-    scope: z.object({
-      projectId: z.string(),
-      accessLevel: z.enum(["all", "scores"]),
-    }),
-  }),
-});
-
-export const LegacyIngestionEvent = z.discriminatedUnion("useS3EventStore", [
-  LegacyIngestionEventFull,
-  LegacyIngestionEventMeta,
-]);
+import { eventTypes } from ".";
+import {
+  BatchActionQuerySchema,
+  BatchActionType,
+} from "../features/batchAction/types";
+import { BatchExportTableName } from "../features/batchExport/types";
 
 export const IngestionEvent = z.object({
   data: z.object({
     type: z.nativeEnum(eventTypes),
     eventBodyId: z.string(),
+    fileKey: z.string().optional(),
   }),
   authCheck: z.object({
     validKey: z.literal(true),
@@ -86,6 +57,19 @@ export const ExperimentCreateEventSchema = z.object({
   runId: z.string(),
   description: z.string().optional(),
 });
+export const DataRetentionProcessingEventSchema = z.object({
+  projectId: z.string(),
+  retention: z.number(),
+});
+export const BatchActionProcessingEventSchema = z.object({
+  projectId: z.string(),
+  actionId: z.string(),
+  query: BatchActionQuerySchema,
+  tableName: z.nativeEnum(BatchExportTableName),
+  cutoffCreatedAt: z.date(),
+  targetId: z.string().optional(),
+  type: z.nativeEnum(BatchActionType),
+});
 
 export type BatchExportJobType = z.infer<typeof BatchExportJobSchema>;
 export type TraceQueueEventType = z.infer<typeof TraceQueueEventSchema>;
@@ -95,7 +79,6 @@ export type DatasetRunItemUpsertEventType = z.infer<
   typeof DatasetRunItemUpsertEventSchema
 >;
 export type EvalExecutionEventType = z.infer<typeof EvalExecutionEvent>;
-export type LegacyIngestionEventType = z.infer<typeof LegacyIngestionEvent>;
 export type IngestionEventQueueType = z.infer<typeof IngestionEvent>;
 export type ExperimentCreateEventType = z.infer<
   typeof ExperimentCreateEventSchema
@@ -103,6 +86,12 @@ export type ExperimentCreateEventType = z.infer<
 export type PostHogIntegrationProcessingEventType = z.infer<
   typeof PostHogIntegrationProcessingEventSchema
 >;
+export type DataRetentionProcessingEventType = z.infer<
+  typeof DataRetentionProcessingEventSchema
+>;
+export type BatchActionProcessingEventType = z.infer<
+  typeof BatchActionProcessingEventSchema
+>;
 
 export enum QueueName {
   TraceUpsert = "trace-upsert", // Ingestion pipeline adds events on each Trace upsert
@@ -112,11 +101,16 @@ export enum QueueName {
   DatasetRunItemUpsert = "dataset-run-item-upsert-queue",
   BatchExport = "batch-export-queue",
   IngestionQueue = "ingestion-queue", // Process single events with S3-merge
-  LegacyIngestionQueue = "legacy-ingestion-queue", // Used for batch processing of Ingestion
+  IngestionSecondaryQueue = "secondary-ingestion-queue", // Separates high priority + high throughput projects from other projects.
   CloudUsageMeteringQueue = "cloud-usage-metering-queue",
   ExperimentCreate = "experiment-create-queue",
   PostHogIntegrationQueue = "posthog-integration-queue",
   PostHogIntegrationProcessingQueue = "posthog-integration-processing-queue",
+  CoreDataS3ExportQueue = "core-data-s3-export-queue",
+  MeteringDataPostgresExportQueue = "metering-data-postgres-export-queue",
+  DataRetentionQueue = "data-retention-queue",
+  DataRetentionProcessingQueue = "data-retention-processing-queue",
+  BatchActionQueue = "batch-action-queue",
 }
 
 export enum QueueJobs {
@@ -126,12 +120,17 @@ export enum QueueJobs {
   DatasetRunItemUpsert = "dataset-run-item-upsert",
   EvaluationExecution = "evaluation-execution-job",
   BatchExportJob = "batch-export-job",
-  LegacyIngestionJob = "legacy-ingestion-job",
   CloudUsageMeteringJob = "cloud-usage-metering-job",
   IngestionJob = "ingestion-job",
+  IngestionSecondaryJob = "secondary-ingestion-job",
   ExperimentCreateJob = "experiment-create-job",
   PostHogIntegrationJob = "posthog-integration-job",
   PostHogIntegrationProcessingJob = "posthog-integration-processing-job",
+  CoreDataS3ExportJob = "core-data-s3-export-job",
+  MeteringDataPostgresExportJob = "metering-data-postgres-export-job",
+  DataRetentionJob = "data-retention-job",
+  DataRetentionProcessingJob = "data-retention-processing-job",
+  BatchActionProcessingJob = "batch-action-processing-job",
 }
 
 export type TQueueJobTypes = {
@@ -171,13 +170,13 @@ export type TQueueJobTypes = {
     payload: BatchExportJobType;
     name: QueueJobs.BatchExportJob;
   };
-  [QueueName.LegacyIngestionQueue]: {
+  [QueueName.IngestionQueue]: {
     timestamp: Date;
     id: string;
-    payload: LegacyIngestionEventType;
-    name: QueueJobs.LegacyIngestionJob;
+    payload: IngestionEventQueueType;
+    name: QueueJobs.IngestionJob;
   };
-  [QueueName.IngestionQueue]: {
+  [QueueName.IngestionSecondaryQueue]: {
     timestamp: Date;
     id: string;
     payload: IngestionEventQueueType;
@@ -195,4 +194,16 @@ export type TQueueJobTypes = {
     payload: PostHogIntegrationProcessingEventType;
     name: QueueJobs.PostHogIntegrationProcessingJob;
   };
+  [QueueName.DataRetentionProcessingQueue]: {
+    timestamp: Date;
+    id: string;
+    payload: DataRetentionProcessingEventType;
+    name: QueueJobs.DataRetentionProcessingJob;
+  };
+  [QueueName.BatchActionQueue]: {
+    timestamp: Date;
+    id: string;
+    payload: BatchActionProcessingEventType;
+    name: QueueJobs.BatchActionProcessingJob;
+  };
 };

packages/shared/src/server/redis/batchActionQueue.ts

@@ -3,30 +3,30 @@ import { QueueName, TQueueJobTypes } from "../queues";
 import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
 import { logger } from "../logger";
 
-export class LegacyIngestionQueue {
+export class BatchActionQueue {
   private static instance: Queue<
-    TQueueJobTypes[QueueName.LegacyIngestionQueue]
+    TQueueJobTypes[QueueName.BatchActionQueue]
   > | null = null;
 
   public static getInstance(): Queue<
-    TQueueJobTypes[QueueName.LegacyIngestionQueue]
+    TQueueJobTypes[QueueName.BatchActionQueue]
   > | null {
-    if (LegacyIngestionQueue.instance) return LegacyIngestionQueue.instance;
+    if (BatchActionQueue.instance) return BatchActionQueue.instance;
 
     const newRedis = createNewRedisInstance({
       enableOfflineQueue: false,
       ...redisQueueRetryOptions,
     });
 
-    LegacyIngestionQueue.instance = newRedis
-      ? new Queue<TQueueJobTypes[QueueName.LegacyIngestionQueue]>(
-          QueueName.LegacyIngestionQueue,
+    BatchActionQueue.instance = newRedis
+      ? new Queue<TQueueJobTypes[QueueName.BatchActionQueue]>(
+          QueueName.BatchActionQueue,
           {
             connection: newRedis,
             defaultJobOptions: {
               removeOnComplete: true,
-              removeOnFail: 500_000,
-              attempts: 5,
+              removeOnFail: 10_000,
+              attempts: 2,
               backoff: {
                 type: "exponential",
                 delay: 5000,
@@ -36,10 +36,10 @@ export class LegacyIngestionQueue {
         )
       : null;
 
-    LegacyIngestionQueue.instance?.on("error", (err) => {
-      logger.error("LegacyIngestionQueue error", err);
+    BatchActionQueue.instance?.on("error", (err) => {
+      logger.error("BatchActionQueue error", err);
     });
 
-    return LegacyIngestionQueue.instance;
+    return BatchActionQueue.instance;
   }
 }

packages/shared/src/server/redis/cloudUsageMeteringQueue.ts

@@ -45,6 +45,7 @@ export class CloudUsageMeteringQueue {
         QueueJobs.CloudUsageMeteringJob,
         {},
         {
+          // Run at minute 5 of every hour (e.g. 1:05, 2:05, 3:05, etc)
           repeat: { pattern: "5 * * * *" },
         },
       );

packages/shared/src/server/redis/coreDataS3ExportQueue.ts

@@ -0,0 +1,60 @@
+import { Queue } from "bullmq";
+import { QueueName, QueueJobs } from "../queues";
+import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
+import { logger } from "../logger";
+import { env } from "../../env";
+
+export class CoreDataS3ExportQueue {
+  private static instance: Queue | null = null;
+
+  public static getInstance(): Queue | null {
+    if (env.LANGFUSE_S3_CORE_DATA_EXPORT_IS_ENABLED !== "true") {
+      return null;
+    }
+
+    if (CoreDataS3ExportQueue.instance) {
+      return CoreDataS3ExportQueue.instance;
+    }
+
+    const newRedis = createNewRedisInstance({
+      enableOfflineQueue: false,
+      ...redisQueueRetryOptions,
+    });
+
+    CoreDataS3ExportQueue.instance = newRedis
+      ? new Queue(QueueName.CoreDataS3ExportQueue, {
+          connection: newRedis,
+          defaultJobOptions: {
+            removeOnComplete: true,
+            removeOnFail: 100,
+            attempts: 5,
+            backoff: {
+              type: "exponential",
+              delay: 5000,
+            },
+          },
+        })
+      : null;
+
+    CoreDataS3ExportQueue.instance?.on("error", (err) => {
+      logger.error("CoreDataS3ExportQueue error", err);
+    });
+
+    if (CoreDataS3ExportQueue.instance) {
+      logger.debug("Scheduling jobs for CoreDataS3ExportQueue");
+      CoreDataS3ExportQueue.instance
+        .add(
+          QueueJobs.CoreDataS3ExportJob,
+          {},
+          {
+            repeat: { pattern: "15 3 * * *" }, // every day at 3:15am
+          },
+        )
+        .catch((err) => {
+          logger.error("Error adding CoreDataS3ExportJob schedule", err);
+        });
+    }
+
+    return CoreDataS3ExportQueue.instance;
+  }
+}

packages/shared/src/server/redis/dataRetentionProcessingQueue.ts

@@ -0,0 +1,40 @@
+import { Queue } from "bullmq";
+import { QueueName } from "../queues";
+import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
+import { logger } from "../logger";
+
+export class DataRetentionProcessingQueue {
+  private static instance: Queue | null = null;
+
+  public static getInstance(): Queue | null {
+    if (DataRetentionProcessingQueue.instance) {
+      return DataRetentionProcessingQueue.instance;
+    }
+
+    const newRedis = createNewRedisInstance({
+      enableOfflineQueue: false,
+      ...redisQueueRetryOptions,
+    });
+
+    DataRetentionProcessingQueue.instance = newRedis
+      ? new Queue(QueueName.DataRetentionProcessingQueue, {
+          connection: newRedis,
+          defaultJobOptions: {
+            removeOnComplete: true,
+            removeOnFail: 10000,
+            attempts: 5,
+            backoff: {
+              type: "exponential",
+              delay: 5000,
+            },
+          },
+        })
+      : null;
+
+    DataRetentionProcessingQueue.instance?.on("error", (err) => {
+      logger.error("DataRetentionProcessingQueue error", err);
+    });
+
+    return DataRetentionProcessingQueue.instance;
+  }
+}

packages/shared/src/server/redis/dataRetentionQueue.ts

@@ -0,0 +1,55 @@
+import { Queue } from "bullmq";
+import { QueueName, QueueJobs } from "../queues";
+import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
+import { logger } from "../logger";
+
+export class DataRetentionQueue {
+  private static instance: Queue | null = null;
+
+  public static getInstance(): Queue | null {
+    if (DataRetentionQueue.instance) {
+      return DataRetentionQueue.instance;
+    }
+
+    const newRedis = createNewRedisInstance({
+      enableOfflineQueue: false,
+      ...redisQueueRetryOptions,
+    });
+
+    DataRetentionQueue.instance = newRedis
+      ? new Queue(QueueName.DataRetentionQueue, {
+          connection: newRedis,
+          defaultJobOptions: {
+            removeOnComplete: true,
+            removeOnFail: 100,
+            attempts: 5,
+            backoff: {
+              type: "exponential",
+              delay: 5000,
+            },
+          },
+        })
+      : null;
+
+    DataRetentionQueue.instance?.on("error", (err) => {
+      logger.error("DataRetentionQueue error", err);
+    });
+
+    if (DataRetentionQueue.instance) {
+      logger.debug("Scheduling jobs for DataRetentionQueue");
+      DataRetentionQueue.instance
+        .add(
+          QueueJobs.DataRetentionJob,
+          {},
+          {
+            repeat: { pattern: "15 3 * * *" }, // every day at 3:15am
+          },
+        )
+        .catch((err) => {
+          logger.error("Error adding DataRetentionQueue schedule", err);
+        });
+    }
+
+    return DataRetentionQueue.instance;
+  }
+}

packages/shared/src/server/redis/getQueue.ts

@@ -5,18 +5,20 @@ import { CloudUsageMeteringQueue } from "./cloudUsageMeteringQueue";
 import { DatasetRunItemUpsertQueue } from "./datasetRunItemUpsert";
 import { EvalExecutionQueue } from "./evalExecutionQueue";
 import { ExperimentCreateQueue } from "./experimentCreateQueue";
-import { IngestionQueue } from "./ingestionQueue";
-import { LegacyIngestionQueue } from "./legacyIngestion";
+import { IngestionQueue, SecondaryIngestionQueue } from "./ingestionQueue";
 import { TraceUpsertQueue } from "./traceUpsert";
 import { TraceDeleteQueue } from "./traceDelete";
 import { ProjectDeleteQueue } from "./projectDelete";
 import { PostHogIntegrationQueue } from "./postHogIntegrationQueue";
 import { PostHogIntegrationProcessingQueue } from "./postHogIntegrationProcessingQueue";
+import { CoreDataS3ExportQueue } from "./coreDataS3ExportQueue";
+import { MeteringDataPostgresExportQueue } from "./meteringDataPostgresExportQueue";
+import { DataRetentionQueue } from "./dataRetentionQueue";
+import { DataRetentionProcessingQueue } from "./dataRetentionProcessingQueue";
+import { BatchActionQueue } from "./batchActionQueue";
 
 export function getQueue(queueName: QueueName): Queue | null {
   switch (queueName) {
-    case QueueName.LegacyIngestionQueue:
-      return LegacyIngestionQueue.getInstance();
     case QueueName.BatchExport:
       return BatchExportQueue.getInstance();
     case QueueName.CloudUsageMeteringQueue:
@@ -39,6 +41,18 @@ export function getQueue(queueName: QueueName): Queue | null {
       return PostHogIntegrationQueue.getInstance();
     case QueueName.PostHogIntegrationProcessingQueue:
       return PostHogIntegrationProcessingQueue.getInstance();
+    case QueueName.IngestionSecondaryQueue:
+      return SecondaryIngestionQueue.getInstance();
+    case QueueName.CoreDataS3ExportQueue:
+      return CoreDataS3ExportQueue.getInstance();
+    case QueueName.MeteringDataPostgresExportQueue:
+      return MeteringDataPostgresExportQueue.getInstance();
+    case QueueName.DataRetentionQueue:
+      return DataRetentionQueue.getInstance();
+    case QueueName.DataRetentionProcessingQueue:
+      return DataRetentionProcessingQueue.getInstance();
+    case QueueName.BatchActionQueue:
+      return BatchActionQueue.getInstance();
     default:
       const exhaustiveCheckDefault: never = queueName;
       throw new Error(`Queue ${queueName} not found`);

packages/shared/src/server/redis/ingestionQueue.ts

@@ -43,3 +43,45 @@ export class IngestionQueue {
     return IngestionQueue.instance;
   }
 }
+
+export class SecondaryIngestionQueue {
+  private static instance: Queue<
+    TQueueJobTypes[QueueName.IngestionSecondaryQueue]
+  > | null = null;
+
+  public static getInstance(): Queue<
+    TQueueJobTypes[QueueName.IngestionSecondaryQueue]
+  > | null {
+    if (SecondaryIngestionQueue.instance)
+      return SecondaryIngestionQueue.instance;
+
+    const newRedis = createNewRedisInstance({
+      enableOfflineQueue: false,
+      ...redisQueueRetryOptions,
+    });
+
+    SecondaryIngestionQueue.instance = newRedis
+      ? new Queue<TQueueJobTypes[QueueName.IngestionSecondaryQueue]>(
+          QueueName.IngestionSecondaryQueue,
+          {
+            connection: newRedis,
+            defaultJobOptions: {
+              removeOnComplete: true,
+              removeOnFail: 100_000,
+              attempts: 5,
+              backoff: {
+                type: "exponential",
+                delay: 5000,
+              },
+            },
+          },
+        )
+      : null;
+
+    SecondaryIngestionQueue.instance?.on("error", (err) => {
+      logger.error("SecondaryIngestionQueue error", err);
+    });
+
+    return SecondaryIngestionQueue.instance;
+  }
+}

packages/shared/src/server/redis/meteringDataPostgresExportQueue.ts

@@ -0,0 +1,63 @@
+import { Queue } from "bullmq";
+import { QueueName, QueueJobs } from "../queues";
+import { createNewRedisInstance, redisQueueRetryOptions } from "./redis";
+import { logger } from "../logger";
+import { env } from "../../env";
+
+export class MeteringDataPostgresExportQueue {
+  private static instance: Queue | null = null;
+
+  public static getInstance(): Queue | null {
+    if (env.LANGFUSE_POSTGRES_METERING_DATA_EXPORT_IS_ENABLED !== "true") {
+      return null;
+    }
+
+    if (MeteringDataPostgresExportQueue.instance) {
+      return MeteringDataPostgresExportQueue.instance;
+    }
+
+    const newRedis = createNewRedisInstance({
+      enableOfflineQueue: false,
+      ...redisQueueRetryOptions,
+    });
+
+    MeteringDataPostgresExportQueue.instance = newRedis
+      ? new Queue(QueueName.MeteringDataPostgresExportQueue, {
+          connection: newRedis,
+          defaultJobOptions: {
+            removeOnComplete: true,
+            removeOnFail: 100,
+            attempts: 5,
+            backoff: {
+              type: "exponential",
+              delay: 5000,
+            },
+          },
+        })
+      : null;
+
+    MeteringDataPostgresExportQueue.instance?.on("error", (err) => {
+      logger.error("MeteringDataPostgresExportQueue error", err);
+    });
+
+    if (MeteringDataPostgresExportQueue.instance) {
+      logger.debug("Scheduling jobs for MeteringDataPostgresExportQueue");
+      MeteringDataPostgresExportQueue.instance
+        .add(
+          QueueJobs.MeteringDataPostgresExportJob,
+          {},
+          {
+            repeat: { pattern: "30 2 * * *" }, // every day at 2:30am UTC
+          },
+        )
+        .catch((err) => {
+          logger.error(
+            "Error adding MeteringDataPostgresExportJob schedule",
+            err,
+          );
+        });
+    }
+
+    return MeteringDataPostgresExportQueue.instance;
+  }
+}

packages/shared/src/server/redis/projectDelete.ts

@@ -27,6 +27,7 @@ export class ProjectDeleteQueue {
               removeOnComplete: true,
               removeOnFail: 100_000,
               attempts: 5,
+              delay: 60_000, // 1 minute
               backoff: {
                 type: "exponential",
                 delay: 5000,

packages/shared/src/server/repositories/clickhouse.ts

@@ -4,14 +4,15 @@ import {
   convertDateToClickhouseDateTime,
 } from "../clickhouse/client";
 import { logger } from "../logger";
-import { instrumentAsync } from "../instrumentation";
+import { getTracer, instrumentAsync } from "../instrumentation";
+import { randomUUID } from "crypto";
+import { getClickhouseEntityType } from "../clickhouse/schemaUtils";
+import { NodeClickHouseClientConfigOptions } from "@clickhouse/client/dist/config";
+import { context, trace } from "@opentelemetry/api";
 import {
   StorageService,
   StorageServiceFactory,
 } from "../services/StorageService";
-import { randomUUID } from "crypto";
-import { getClickhouseEntityType } from "../clickhouse/schemaUtils";
-import { NodeClickHouseClientConfigOptions } from "@clickhouse/client/dist/config";
 
 let s3StorageServiceClient: StorageService;
 
@@ -40,11 +41,8 @@ export async function upsertClickhouse<
     // https://opentelemetry.io/docs/specs/semconv/database/database-spans/
     span.setAttribute("ch.query.table", opts.table);
 
-    const s3Client = getS3StorageServiceClient(
-      env.LANGFUSE_S3_EVENT_UPLOAD_BUCKET,
-    );
     await Promise.all(
-      opts.records.map((record) => {
+      opts.records.map(async (record) => {
         // drop trailing s and pretend it's always a create.
         // Only applicable to scores and traces.
         let eventType = `${opts.table.slice(0, -1)}-create`;
@@ -52,17 +50,38 @@ export async function upsertClickhouse<
           // @ts-ignore - If it's an observation we now that `type` is a string
           eventType = `${record["type"].toLowerCase()}-create`;
         }
-        s3Client.uploadJson(
-          `${env.LANGFUSE_S3_EVENT_UPLOAD_PREFIX}${record.project_id}/${getClickhouseEntityType(eventType)}/${record.id}/${randomUUID()}.json`,
-          [
+
+        const eventId = randomUUID();
+        const bucketPath = `${env.LANGFUSE_S3_EVENT_UPLOAD_PREFIX}${record.project_id}/${getClickhouseEntityType(eventType)}/${record.id}/${eventId}.json`;
+
+        // Write new file directly to ClickHouse. We don't use the ClickHouse writer here as we expect more limited traffic
+        // and are not worried that much about latency.
+        await clickhouseClient().insert({
+          table: "event_log",
+          values: [
             {
               id: randomUUID(),
-              timestamp: new Date().toISOString(),
-              type: eventType,
-              body: opts.eventBodyMapper(record),
+              project_id: record.project_id,
+              entity_type: getClickhouseEntityType(eventType),
+              entity_id: record.id,
+              event_id: eventId,
+              bucket_name: env.LANGFUSE_S3_EVENT_UPLOAD_BUCKET,
+              bucket_path: bucketPath,
             },
           ],
-        );
+          format: "JSONEachRow",
+        });
+
+        return getS3StorageServiceClient(
+          env.LANGFUSE_S3_EVENT_UPLOAD_BUCKET,
+        ).uploadJson(bucketPath, [
+          {
+            id: eventId,
+            timestamp: new Date().toISOString(),
+            type: eventType,
+            body: opts.eventBodyMapper(record),
+          },
+        ]);
       }),
     );
 
@@ -101,6 +120,64 @@ export async function upsertClickhouse<
   });
 }
 
+export async function* queryClickhouseStream<T>(opts: {
+  query: string;
+  params?: Record<string, unknown> | undefined;
+  clickhouseConfigs?: NodeClickHouseClientConfigOptions;
+}): AsyncGenerator<T> {
+  const tracer = getTracer("clickhouse-query-stream");
+  const span = tracer.startSpan("clickhouse-query-stream");
+
+  try {
+    const res = await context.with(
+      trace.setSpan(context.active(), span),
+      async () => {
+        // https://opentelemetry.io/docs/specs/semconv/database/database-spans/
+        span.setAttribute("ch.query.text", opts.query);
+
+        const res = await clickhouseClient(opts.clickhouseConfigs).query({
+          query: opts.query,
+          format: "JSONEachRow",
+          query_params: opts.params,
+        });
+        // same logic as for prisma. we want to see queries in development
+        if (env.NODE_ENV === "development") {
+          logger.info(`clickhouse:query ${res.query_id} ${opts.query}`);
+        }
+
+        span.setAttribute("ch.queryId", res.query_id);
+
+        // add summary headers to the span. Helps to tune performance
+        const summaryHeader = res.response_headers["x-clickhouse-summary"];
+        if (summaryHeader) {
+          try {
+            const summary = Array.isArray(summaryHeader)
+              ? JSON.parse(summaryHeader[0])
+              : JSON.parse(summaryHeader);
+            for (const key in summary) {
+              span.setAttribute(`ch.${key}`, summary[key]);
+            }
+          } catch (error) {
+            logger.debug(
+              `Failed to parse clickhouse summary header ${summaryHeader}`,
+              error,
+            );
+          }
+        }
+        return res;
+      },
+    );
+
+    for await (const rows of res.stream<T>()) {
+      for (const row of rows) {
+        yield row.json();
+      }
+    }
+  } finally {
+    span.end();
+  }
+}
+
 export async function queryClickhouse<T>(opts: {
   query: string;
   params?: Record<string, unknown> | undefined;

packages/shared/src/server/repositories/dashboards.ts

@@ -59,7 +59,6 @@ export const getObservationsCostGroupedByName = async (
   const appliedFilter = chFilter.apply();
 
   const hasTraceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
-  // TODO: Validate whether we can filter traces on timestamp here.
 
   const query = `
     SELECT 
@@ -68,9 +67,10 @@ export const getObservationsCostGroupedByName = async (
       sumMap(usage_details)['total'] as sum_usage_details
     FROM observations o FINAL ${hasTraceFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
     WHERE project_id = {projectId: String}
-    AND ${appliedFilter.query}
+    ${appliedFilter.query ? `AND ${appliedFilter.query}` : ""}
     GROUP BY provided_model_name
     ORDER BY sumMap(cost_details)['total'] DESC
+    LIMIT 50
     `;
 
   const result = await queryClickhouse<{
@@ -145,23 +145,26 @@ export const getScoreAggregate = async (
 export const groupTracesByTime = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
   ).apply();
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "timestamp",
+  );
+
   const query = `
     SELECT 
-      ${selectTimeseriesColumn(groupBy, "timestamp", "timestamp")},
+      ${selectTimeseriesColumn(bucketSizeInSeconds, "timestamp", "timestamp")},
       count(*) as count
     FROM traces t FINAL
     WHERE project_id = {projectId: String}
     AND ${chFilter.query}
     GROUP BY timestamp
-    ${orderByTimeSeries(groupBy, "timestamp")}
+    ${orderByQuery}
     `;
-
   const result = await queryClickhouse<{
     timestamp: string;
     count: string;
@@ -170,11 +173,12 @@ export const groupTracesByTime = async (
     params: {
       projectId,
       ...chFilter.params,
+      ...orderByParams,
     },
   });
 
   return result.map((row) => ({
-    timestamp: new Date(row.timestamp),
+    timestamp: parseClickhouseUTCDateTimeFormat(row.timestamp),
     countTraceId: Number(row.count),
   }));
 };
@@ -182,7 +186,6 @@ export const groupTracesByTime = async (
 export const getObservationUsageByTime = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -200,9 +203,14 @@ export const getObservationUsageByTime = async (
       ) as DateTimeFilter | undefined)
     : undefined;
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "start_time",
+  );
+
   const query = `
     SELECT 
-      ${selectTimeseriesColumn(groupBy, "start_time", "start_time")},
+      ${selectTimeseriesColumn(bucketSizeInSeconds, "start_time", "start_time")},
       sumMap(usage_details) as units,
       sumMap(cost_details) as cost, 
       provided_model_name
@@ -212,7 +220,7 @@ export const getObservationUsageByTime = async (
     AND ${appliedFilter.query}
     ${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
     GROUP BY start_time, provided_model_name
-    ${orderByTimeSeries(groupBy, "start_time")}
+    ${orderByQuery}
     `;
 
   const result = await queryClickhouse<{
@@ -225,6 +233,7 @@ export const getObservationUsageByTime = async (
     params: {
       projectId,
       ...appliedFilter.params,
+      ...orderByParams,
       ...(timeFilter
         ? { traceTimestamp: convertDateToClickhouseDateTime(timeFilter.value) }
         : {}),
@@ -232,7 +241,7 @@ export const getObservationUsageByTime = async (
   });
 
   return result.map((row) => ({
-    start_time: new Date(row.start_time),
+    start_time: parseClickhouseUTCDateTimeFormat(row.start_time),
     units: Object.fromEntries(
       Object.entries(row.units ?? {}).map(([key, value]) => [
         key,
@@ -271,12 +280,15 @@ export const getDistinctModels = async (
 
   // No need for final as duplicates are caught by distinct anyway.
   const query = `
-    SELECT distinct(provided_model_name) as model
+    SELECT distinct(provided_model_name) as model, count(*) as count
     FROM observations o
     ${tracesFilter ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
     WHERE project_id = {projectId: String}
     AND ${appliedFilter.query}
     ${timeFilter ? `AND t.timestamp >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
+    GROUP BY provided_model_name
+    ORDER BY count(*) DESC
+    LIMIT 1000
     `;
 
   const result = await queryClickhouse<{ model: string }>({
@@ -296,7 +308,6 @@ export const getDistinctModels = async (
 export const getScoresAggregateOverTime = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -307,9 +318,14 @@ export const getScoresAggregateOverTime = async (
   const traceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
   // TODO: Validate whether we can filter traces on timestamp here.
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "timestamp",
+  );
+
   const query = `
   SELECT 
-    ${selectTimeseriesColumn(groupBy, "timestamp", "timestamp")},
+    ${selectTimeseriesColumn(bucketSizeInSeconds, "timestamp", "timestamp")},
     name,
     data_type,
     source,
@@ -324,7 +340,7 @@ export const getScoresAggregateOverTime = async (
     name,
     data_type,
     source
-  ${orderByTimeSeries(groupBy, "timestamp")};
+  ${orderByQuery};
 `;
 
   const result = await queryClickhouse<{
@@ -338,11 +354,12 @@ export const getScoresAggregateOverTime = async (
     params: {
       projectId,
       ...appliedFilter.params,
+      ...orderByParams,
     },
   });
 
   return result.map((row) => ({
-    scoreTimestamp: new Date(row.timestamp),
+    scoreTimestamp: parseClickhouseUTCDateTimeFormat(row.timestamp),
     scoreName: row.name,
     scoreDataType: row.data_type,
     scoreSource: row.source,
@@ -418,7 +435,7 @@ export const getObservationLatencies = async (
   // Skipping FINAL here, as the quantiles are approximate to begin with.
   const query = `
     SELECT
-      quantiles(0.5, 0.9, 0.95, 0.99)(date_diff('milliseconds', o.start_time, o.end_time)) as quantiles,
+      quantiles(0.5, 0.9, 0.95, 0.99)(date_diff('millisecond', o.start_time, o.end_time)) as quantiles,
       name
     FROM observations o
     ${chFilter.find((f) => f.clickhouseTable === "traces") ? "LEFT JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
@@ -465,7 +482,7 @@ export const getTracesLatencies = async (
       select o.trace_id,
              t.name,
              o.project_id,
-             date_diff('milliseconds', min(o.start_time), coalesce(max(o.end_time), max(o.start_time))) as duration
+             date_diff('millisecond', min(o.start_time), coalesce(max(o.end_time), max(o.start_time))) as duration
       FROM traces t 
       JOIN observations o
       ON o.trace_id = t.id AND o.project_id = t.project_id
@@ -506,7 +523,6 @@ export const getTracesLatencies = async (
 export const getModelLatenciesOverTime = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -516,25 +532,33 @@ export const getModelLatenciesOverTime = async (
 
   const traceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "start_time_bucket",
+  );
+
   // Skipping FINAL here, as the quantiles are approximate to begin with.
   const query = `
   SELECT 
-    ${selectTimeseriesColumn(groupBy, "o.start_time", "start_time_bucket")},
+    ${selectTimeseriesColumn(bucketSizeInSeconds, "o.start_time", "start_time_bucket")},
     provided_model_name,
-    quantiles(0.5, 0.75, 0.9, 0.95, 0.99)(date_diff('milliseconds', o.start_time, o.end_time)) as quantiles
+    quantiles(0.5, 0.75, 0.9, 0.95, 0.99)(date_diff('millisecond', o.start_time, o.end_time)) as quantiles
   FROM observations o
   ${traceFilter ? "JOIN traces t ON o.trace_id = t.id AND o.project_id = t.project_id" : ""}
   WHERE project_id = {projectId: String}
   AND ${appliedFilter.query}
   GROUP BY provided_model_name, start_time_bucket
-  ${orderByTimeSeries(groupBy, "start_time_bucket")};
+  ${orderByQuery};
 `;
 
   const result = await queryClickhouse<{
     start_time_bucket: string;
     provided_model_name: string;
     quantiles: string[];
-  }>({ query, params: { projectId, ...appliedFilter.params } });
+  }>({
+    query,
+    params: { projectId, ...appliedFilter.params, ...orderByParams },
+  });
 
   return result.map((row) => ({
     p50: Number(row.quantiles[0]) / 1000,
@@ -543,14 +567,13 @@ export const getModelLatenciesOverTime = async (
     p95: Number(row.quantiles[3]) / 1000,
     p99: Number(row.quantiles[4]) / 1000,
     model: row.provided_model_name,
-    start_time: new Date(row.start_time_bucket),
+    start_time: parseClickhouseUTCDateTimeFormat(row.start_time_bucket),
   }));
 };
 
 export const getNumericScoreTimeSeries = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -559,9 +582,14 @@ export const getNumericScoreTimeSeries = async (
 
   const traceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "score_timestamp",
+  );
+
   const query = `
     SELECT
-    ${selectTimeseriesColumn(groupBy, "s.timestamp", "score_timestamp")},
+    ${selectTimeseriesColumn(bucketSizeInSeconds, "s.timestamp", "score_timestamp")},
     s.name as score_name,
     AVG(s.value) as avg_value
     FROM scores s final
@@ -569,11 +597,11 @@ export const getNumericScoreTimeSeries = async (
     WHERE s.project_id = {projectId: String}
     ${chFilterRes?.query ? `AND ${chFilterRes.query}` : ""}
     GROUP BY score_name, score_timestamp
-    ${orderByTimeSeries(groupBy, "score_timestamp")}
+    ${orderByQuery}
   `;
 
-  return queryClickhouse<{
-    score_timestamp: Date;
+  const result = await queryClickhouse<{
+    score_timestamp: string;
     score_name: string;
     avg_value: number;
   }>({
@@ -581,14 +609,20 @@ export const getNumericScoreTimeSeries = async (
     params: {
       projectId,
       ...(chFilterRes ? chFilterRes.params : {}),
+      ...orderByParams,
     },
   });
+
+  return result.map((row) => ({
+    scoreTimestamp: parseClickhouseUTCDateTimeFormat(row.score_timestamp),
+    scoreName: row.score_name,
+    avgValue: Number(row.avg_value),
+  }));
 };
 
 export const getCategoricalScoreTimeSeries = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc | undefined,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -597,9 +631,14 @@ export const getCategoricalScoreTimeSeries = async (
 
   const traceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "score_timestamp",
+  );
+
   const query = `
     SELECT
-    ${groupBy ? selectTimeseriesColumn(groupBy, "s.timestamp", "score_timestamp") + ", " : ""}
+    ${bucketSizeInSeconds ? selectTimeseriesColumn(bucketSizeInSeconds, "s.timestamp", "score_timestamp") + ", " : ""}
     s.name as score_name,
     s.data_type as score_data_type,
     s.source as score_source,
@@ -609,12 +648,12 @@ export const getCategoricalScoreTimeSeries = async (
     ${traceFilter ? "JOIN traces t ON s.trace_id = t.id AND s.project_id = t.project_id" : ""}
     WHERE s.project_id = {projectId: String}
     ${chFilterRes?.query ? `AND ${chFilterRes.query}` : ""}
-    GROUP BY score_name, score_data_type, score_source, score_value ${groupBy ? ", score_timestamp" : ""}
-    ${groupBy ? orderByTimeSeries(groupBy, "score_timestamp") : ""}
+    GROUP BY score_name, score_data_type, score_source, score_value ${bucketSizeInSeconds ? ", score_timestamp" : ""}
+      ${orderByQuery}
   `;
 
-  return queryClickhouse<{
-    score_timestamp?: Date;
+  const result = await queryClickhouse<{
+    score_timestamp?: string;
     score_name: string;
     score_data_type: string;
     score_source: string;
@@ -625,14 +664,25 @@ export const getCategoricalScoreTimeSeries = async (
     params: {
       projectId,
       ...(chFilterRes ? chFilterRes.params : {}),
+      ...orderByParams,
     },
   });
+
+  return result.map((row) => ({
+    scoreTimestamp: row.score_timestamp
+      ? parseClickhouseUTCDateTimeFormat(row.score_timestamp)
+      : undefined,
+    scoreName: row.score_name,
+    scoreDataType: row.score_data_type,
+    scoreSource: row.score_source,
+    scoreValue: row.score_value,
+    count: Number(row.count),
+  }));
 };
 
 export const getObservationsStatusTimeSeries = async (
   projectId: string,
   filter: FilterState,
-  groupBy: DateTrunc | undefined,
 ) => {
   const chFilter = new FilterList(
     createFilterFromFilterState(filter, dashboardColumnDefinitions),
@@ -641,9 +691,14 @@ export const getObservationsStatusTimeSeries = async (
 
   const traceFilter = chFilter.find((f) => f.clickhouseTable === "traces");
 
+  const [orderByQuery, orderByParams, bucketSizeInSeconds] = orderByTimeSeries(
+    filter,
+    "start_time_bucket",
+  );
+
   const query = `
     SELECT 
-      ${groupBy ? selectTimeseriesColumn(groupBy, "o.start_time", "start_time_bucket") + ", " : ""}
+      ${bucketSizeInSeconds ? selectTimeseriesColumn(bucketSizeInSeconds, "o.start_time", "start_time_bucket") + ", " : ""}
       count(*) as observation_count,
       level as level
     FROM observations o
@@ -651,8 +706,8 @@ export const getObservationsStatusTimeSeries = async (
     WHERE project_id = {projectId: String}
     AND o.level IS NOT NULL
     AND ${chFilterRes?.query}
-    GROUP BY level ${groupBy ? ", start_time_bucket" : ""}
-    ${groupBy ? orderByTimeSeries(groupBy, "start_time_bucket") : ""}
+    GROUP BY level ${bucketSizeInSeconds ? ", start_time_bucket" : ""}
+    ${orderByQuery}
   `;
 
   const result = await queryClickhouse<{
@@ -661,7 +716,11 @@ export const getObservationsStatusTimeSeries = async (
     level: string;
   }>({
     query,
-    params: { projectId, ...(chFilterRes ? chFilterRes.params : {}) },
+    params: {
+      projectId,
+      ...(chFilterRes ? chFilterRes.params : {}),
+      ...orderByParams,
+    },
   });
 
   return result.map((row) => ({
@@ -673,61 +732,70 @@ export const getObservationsStatusTimeSeries = async (
   }));
 };
 
-const orderByTimeSeries = (dateTrunc: DateTrunc, col: string) => {
-  let interval;
-  switch (dateTrunc) {
-    case "year":
-      interval = "toIntervalYear(1)";
-      break;
-    case "month":
-      interval = "toIntervalMonth(1)";
-      break;
-    case "week":
-      interval = "toIntervalWeek(1)";
-      break;
-    case "day":
-      interval = "toIntervalDay(1)";
-      break;
-    case "hour":
-      interval = "toIntervalHour(1)";
-      break;
-    case "minute":
-      interval = "toIntervalMinute(1)";
-      break;
-    default:
-      return undefined;
+export const orderByTimeSeries = (
+  filter: FilterState,
+  col: string,
+): [string, { fromTime: number; toTime: number }, number] => {
+  const potentialBucketSizesSeconds = [
+    5, 10, 30, 60, 300, 600, 1800, 3600, 18000, 36000, 86400, 604800, 2592000,
+  ];
+
+  // Calculate time difference in seconds
+  const [from, to] = extractFromAndToTimestampsFromFilter(filter);
+
+  if (!from || !to) {
+    throw new Error("Time Filter is required for time series queries");
   }
 
-  return `ORDER BY ${col} ASC WITH FILL STEP ${interval}`;
+  const fromDate = new Date(from.value as Date);
+  const toDate = new Date(to.value as Date);
+
+  const diffInSeconds = Math.abs(toDate.getTime() - fromDate.getTime()) / 1000;
+
+  // choose the bucket size that is the closest to the desired number of buckets
+  const bucketSizeInSeconds = potentialBucketSizesSeconds.reduce(
+    (closest, size) => {
+      const diffFromDesiredBuckets = Math.abs(diffInSeconds / size - 50);
+      return diffFromDesiredBuckets < closest.diffFromDesiredBuckets
+        ? { size, diffFromDesiredBuckets }
+        : closest;
+    },
+    { size: 0, diffFromDesiredBuckets: Infinity },
+  ).size;
+
+  // Convert to interval string
+  const interval = `toIntervalSecond(${bucketSizeInSeconds})`;
+
+  return [
+    `ORDER BY ${col} ASC 
+    WITH FILL
+    FROM toStartOfInterval(toDateTime({fromTime: DateTime64(3)}), INTERVAL ${bucketSizeInSeconds} SECOND)
+    TO toDateTime({toTime: DateTime64(3)}) + INTERVAL ${bucketSizeInSeconds} SECOND
+    STEP ${interval}`,
+    { fromTime: fromDate.getTime(), toTime: toDate.getTime() },
+    bucketSizeInSeconds,
+  ];
 };
 
-const selectTimeseriesColumn = (
-  dateTrunc: DateTrunc,
+export const selectTimeseriesColumn = (
+  bucketSizeInSeconds: number,
   col: string,
   as: String,
 ) => {
-  let interval;
-  switch (dateTrunc) {
-    case "year":
-      interval = "toStartOfYear";
-      break;
-    case "month":
-      interval = "toStartOfMonth";
-      break;
-    case "week":
-      interval = "toStartOfWeek";
-      break;
-    case "day":
-      interval = "toStartOfDay";
-      break;
-    case "hour":
-      interval = "toStartOfHour";
-      break;
-    case "minute":
-      interval = "toStartOfMinute";
-      break;
-    default:
-      return undefined;
-  }
-  return `${interval}(${col}) as ${as}`;
+  return `toStartOfInterval(${col}, INTERVAL ${bucketSizeInSeconds} SECOND) as ${as}`;
+};
+
+export const extractFromAndToTimestampsFromFilter = (filter?: FilterState) => {
+  if (!filter)
+    throw new Error("Time Filter is required for time series queries");
+
+  const fromTimestamp = filter.filter(
+    (f) => f.type === "datetime" && (f.operator === ">" || f.operator === ">="),
+  );
+
+  const toTimestamp = filter.filter(
+    (f) => f.type === "datetime" && (f.operator === "<" || f.operator === "<="),
+  );
+
+  return [fromTimestamp[0], toTimestamp[0]];
 };

packages/shared/src/server/repositories/definitions.ts

@@ -158,6 +158,30 @@ export const scoreRecordInsertSchema = scoreRecordBaseSchema.extend({
 });
 export type ScoreRecordInsertType = z.infer<typeof scoreRecordInsertSchema>;
 
+export const eventLogRecordBaseSchema = z.object({
+  id: z.string(),
+  project_id: z.string(),
+  entity_type: z.string(),
+  entity_id: z.string(),
+  // event_id is nullable to be compatible with legacy queue events.
+  // It still allows us to delete things by prefix, but requires an additional list call.
+  event_id: z.string().nullable(),
+  bucket_name: z.string(),
+  bucket_path: z.string(),
+});
+export const eventLogRecordReadSchema = eventLogRecordBaseSchema.extend({
+  created_at: clickhouseStringDateSchema,
+  updated_at: clickhouseStringDateSchema,
+});
+export type EventLogRecordReadType = z.infer<typeof eventLogRecordReadSchema>;
+export const eventLogRecordInsertSchema = eventLogRecordBaseSchema.extend({
+  created_at: z.number(),
+  updated_at: z.number(),
+});
+export type EventLogRecordInsertType = z.infer<
+  typeof eventLogRecordInsertSchema
+>;
+
 export const convertTraceReadToInsert = (
   record: TraceRecordReadType,
 ): TraceRecordInsertType => {

packages/shared/src/server/repositories/eventLog.ts

@@ -0,0 +1,106 @@
+import {
+  commandClickhouse,
+  queryClickhouse,
+  queryClickhouseStream,
+} from "./clickhouse";
+import { EventLogRecordReadType } from "./definitions";
+import { convertDateToClickhouseDateTime } from "../clickhouse/client";
+
+export const getEventLogByProjectAndEntityId = async (
+  projectId: string,
+  entityType: string,
+  entityId: string,
+): Promise<EventLogRecordReadType[]> => {
+  const query = `
+    select *
+    from event_log
+    where project_id = {projectId: String}
+    and entity_type = {entityType: String}
+    and entity_id = {entityId: String}
+  `;
+
+  return queryClickhouse<EventLogRecordReadType>({
+    query,
+    params: {
+      projectId,
+      entityType,
+      entityId,
+    },
+  });
+};
+
+export const getEventLogByProjectId = (
+  projectId: string,
+): AsyncGenerator<EventLogRecordReadType> => {
+  const query = `
+    select *
+    from event_log
+    where project_id = {projectId: String}
+  `;
+
+  return queryClickhouseStream<EventLogRecordReadType>({
+    query,
+    params: {
+      projectId,
+    },
+  });
+};
+
+export const getEventLogByProjectIdBeforeDate = (
+  projectId: string,
+  beforeDate: Date,
+): AsyncGenerator<EventLogRecordReadType> => {
+  const query = `
+        select *
+        from event_log
+        where project_id = {projectId: String}
+        and created_at <= {beforeDate: DateTime64(3)}
+    `;
+
+  return queryClickhouseStream<EventLogRecordReadType>({
+    query,
+    params: {
+      projectId,
+      beforeDate: convertDateToClickhouseDateTime(beforeDate),
+    },
+  });
+};
+
+export const deleteEventLogByProjectId = async (
+  projectId: string,
+): Promise<void> => {
+  const query = `
+    DELETE FROM event_log
+    WHERE project_id = {projectId: String};
+  `;
+  await commandClickhouse({
+    query: query,
+    params: {
+      projectId,
+    },
+    clickhouseConfigs: {
+      request_timeout: 120_000, // 2 minutes
+    },
+  });
+};
+
+export const deleteEventLogByProjectIdBeforeDate = async (
+  projectId: string,
+  beforeDate: Date,
+): Promise<void> => {
+  const query = `
+    DELETE FROM event_log
+    WHERE project_id = {projectId: String}
+    AND created_at <= {beforeDate: DateTime64(3)};
+  `;
+  await commandClickhouse({
+    query: query,
+    params: {
+      projectId,
+      beforeDate: convertDateToClickhouseDateTime(beforeDate),
+    },
+    clickhouseConfigs: {
+      request_timeout: 120_000, // 2 minutes
+    },
+  });
+};

packages/shared/src/server/repositories/index.ts

@@ -9,3 +9,4 @@ export * from "./observations_converters";
 export * from "./clickhouse";
 export * from "./constants";
 export * from "./trace-sessions";
+export * from "./eventLog";

packages/shared/src/server/repositories/observations.ts

@@ -2,6 +2,7 @@ import {
   commandClickhouse,
   parseClickhouseUTCDateTimeFormat,
   queryClickhouse,
+  queryClickhouseStream,
   upsertClickhouse,
 } from "./clickhouse";
 import { ObservationLevel } from "@prisma/client";
@@ -414,7 +415,7 @@ export const getObservationsTable = async (
   const observationRecords = await getObservationsTableInternal<
     Omit<
       ObservationsTableQueryResult,
-      "trace_tags" | "trace_name" | "trace_user_id" | "type"
+      "trace_tags" | "trace_name" | "trace_user_id"
     >
   >({
     ...opts,
@@ -431,7 +432,7 @@ export const getObservationsTable = async (
   return observationRecords.map((o) => {
     const trace = traces.find((t) => t.id === o.trace_id);
     return {
-      ...convertObservationToView({ ...o, type: "GENERATION" }),
+      ...convertObservationToView(o),
       latency: o.latency ? Number(o.latency) / 1000 : null,
       timeToFirstToken: o.time_to_first_token
         ? Number(o.time_to_first_token) / 1000
@@ -449,7 +450,7 @@ export const getObservationsTableWithModelData = async (
   const observationRecords = await getObservationsTableInternal<
     Omit<
       ObservationsTableQueryResult,
-      "trace_tags" | "trace_name" | "trace_user_id" | "type"
+      "trace_tags" | "trace_name" | "trace_user_id"
     >
   >({
     ...opts,
@@ -490,7 +491,7 @@ export const getObservationsTableWithModelData = async (
     const trace = traces.find((t) => t.id === o.trace_id);
     const model = models.find((m) => m.id === o.internal_model_id);
     return {
-      ...convertObservationToView({ ...o, type: "GENERATION" }),
+      ...convertObservationToView(o),
       latency: o.latency ? Number(o.latency) / 1000 : null,
       timeToFirstToken: o.time_to_first_token
         ? Number(o.time_to_first_token) / 1000
@@ -517,6 +518,7 @@ const getObservationsTableInternal = async <T>(
       ? "count(*) as count"
       : `
         o.id as id,
+        o.type as type,
         o.project_id as "project_id",
         o.name as name,
         o."model_parameters" as model_parameters,
@@ -536,9 +538,12 @@ const getObservationsTableInternal = async <T>(
         o.updated_at as "updated_at",
         o.provided_model_name as "provided_model_name",
         o.total_cost as "total_cost",
+        o.prompt_id as "prompt_id",
+        o.prompt_name as "prompt_name",
+        o.prompt_version as "prompt_version",
         internal_model_id as "internal_model_id",
-        if(isNull(end_time), NULL, date_diff('milliseconds', start_time, end_time)) as latency,
-        if(isNull(completion_start_time), NULL,  date_diff('milliseconds', start_time, completion_start_time)) as "time_to_first_token"`;
+        if(isNull(end_time), NULL, date_diff('millisecond', start_time, end_time)) as latency,
+        if(isNull(completion_start_time), NULL,  date_diff('millisecond', start_time, completion_start_time)) as "time_to_first_token"`;
 
   const { projectId, filter, selectIOAndMetadata, limit, offset, orderBy } =
     opts;
@@ -677,7 +682,7 @@ const getObservationsTableInternal = async <T>(
         ${traceTableFilter.length > 0 || orderByTraces || search.query ? "LEFT JOIN traces t FINAL ON t.id = o.trace_id AND t.project_id = o.project_id" : ""}
         ${hasScoresFilter ? `LEFT JOIN scores_avg AS s_avg ON s_avg.trace_id = o.trace_id and s_avg.observation_id = o.id` : ""}
       WHERE ${appliedObservationsFilter.query}
-        AND o.type = 'GENERATION'
+        
         ${timeFilter && (traceTableFilter.length > 0 || orderByTraces) ? `AND t.timestamp > {tracesTimestampFilter: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
         ${search.query}
       ${chOrderBy}
@@ -747,6 +752,50 @@ export const getObservationsGroupedByModel = async (
   return res.map((r) => ({ model: r.name }));
 };
 
+export const getObservationsGroupedByModelId = async (
+  projectId: string,
+  filter: FilterState,
+) => {
+  const observationsFilter = new FilterList([
+    new StringFilter({
+      clickhouseTable: "observations",
+      field: "project_id",
+      operator: "=",
+      value: projectId,
+      tablePrefix: "o",
+    }),
+  ]);
+
+  observationsFilter.push(
+    ...createFilterFromFilterState(
+      filter,
+      observationsTableUiColumnDefinitions,
+    ),
+  );
+
+  const appliedObservationsFilter = observationsFilter.apply();
+
+  // We mainly use queries like this to retrieve filter options.
+  // Therefore, we can skip final as some inaccuracy in count is acceptable.
+  const query = `
+    SELECT o.internal_model_id as modelId
+    FROM observations o
+    WHERE ${appliedObservationsFilter.query}
+    AND o.type = 'GENERATION'
+    GROUP BY o.internal_model_id
+    ORDER BY count() DESC
+    LIMIT 1000;
+  `;
+
+  const res = await queryClickhouse<{ modelId: string }>({
+    query,
+    params: {
+      ...appliedObservationsFilter.params,
+    },
+  });
+  return res.map((r) => ({ modelId: r.modelId }));
+};
+
 export const getObservationsGroupedByName = async (
   projectId: string,
   filter: FilterState,
@@ -859,6 +908,7 @@ export const getObservationsGroupedByPromptName = async (
 
 export const getCostForTraces = async (
   projectId: string,
+  timestamp: Date,
   traceIds: string[],
 ) => {
   // Wrapping the query in a CTE allows us to skip FINAL which allows Clickhouse to use skip indexes.
@@ -868,6 +918,7 @@ export const getCostForTraces = async (
       FROM observations o
       WHERE o.project_id = {projectId: String}
       AND o.trace_id IN ({traceIds: Array(String)})
+      AND o.start_time >= {timestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}
       ORDER BY o.event_ts DESC
       LIMIT 1 BY o.id, o.project_id
     )
@@ -881,6 +932,7 @@ export const getCostForTraces = async (
     params: {
       projectId,
       traceIds,
+      timestamp: convertDateToClickhouseDateTime(timestamp),
     },
   });
   return res.length > 0 ? Number(res[0].total_cost) : undefined;
@@ -923,6 +975,27 @@ export const deleteObservationsByProjectId = async (projectId: string) => {
   });
 };
 
+export const deleteObservationsOlderThanDays = async (
+  projectId: string,
+  days: number,
+) => {
+  const query = `
+    DELETE FROM observations
+    WHERE project_id = {projectId: String}
+    AND start_time < now() - INTERVAL {numDays: Int} DAYS;
+  `;
+  await commandClickhouse({
+    query: query,
+    params: {
+      projectId,
+      numDays: days,
+    },
+    clickhouseConfigs: {
+      request_timeout: 120_000, // 2 minutes
+    },
+  });
+};
+
 export const getObservationsWithPromptName = async (
   projectId: string,
   promptNames: string[],
@@ -963,7 +1036,7 @@ export const getObservationMetricsForPrompts = async (
                   end_time,
                   usage_details,
                   cost_details,
-                  dateDiff('milliseconds', start_time, end_time) AS latency_ms
+                  dateDiff('millisecond', start_time, end_time) AS latency_ms
               FROM observations
               FINAL
               WHERE (type = 'GENERATION') 
@@ -1026,7 +1099,7 @@ export const getLatencyAndTotalCostForObservations = async (
     SELECT
         id,
         cost_details['total'] AS total_cost,
-        dateDiff('milliseconds', start_time, end_time) AS latency_ms
+        dateDiff('millisecond', start_time, end_time) AS latency_ms
     FROM observations FINAL 
     WHERE project_id = {projectId: String} 
     AND id IN ({observationIds: Array(String)})
@@ -1058,7 +1131,7 @@ export const getLatencyAndTotalCostForObservationsByTraces = async (
     SELECT
         trace_id,
         sumMap(cost_details)['total'] AS total_cost,
-        dateDiff('milliseconds', min(start_time), max(end_time)) AS latency_ms
+        dateDiff('millisecond', min(start_time), max(end_time)) AS latency_ms
     FROM observations FINAL
     WHERE project_id = {projectId: String} 
     AND trace_id IN ({traceIds: Array(String)})
@@ -1167,23 +1240,23 @@ export const getTraceIdsForObservations = async (
   }));
 };
 
-export const getGenerationsForPostHog = async (
+export const getGenerationsForPostHog = async function* (
   projectId: string,
   minTimestamp: Date,
   maxTimestamp: Date,
-) => {
+) {
   const query = `
     SELECT
       o.name as name,
       o.start_time as start_time,
       o.id as id,
       o.total_cost as total_cost,
-      if(isNull(completion_start_time), NULL, date_diff('milliseconds', start_time, completion_start_time)) as time_to_first_token,
+      if(isNull(completion_start_time), NULL, date_diff('millisecond', start_time, completion_start_time)) as time_to_first_token,
       o.usage_details['total'] as input_tokens,
       o.usage_details['output'] as output_tokens,
       o.cost_details['total'] as total_tokens,
       o.project_id as project_id,
-      if(isNull(end_time), NULL, date_diff('milliseconds', start_time, end_time) / 1000) as latency,
+      if(isNull(end_time), NULL, date_diff('millisecond', start_time, end_time) / 1000) as latency,
       o.provided_model_name as model,
       o.level as level,
       o.version as version,
@@ -1205,7 +1278,7 @@ export const getGenerationsForPostHog = async (
     AND o.type = 'GENERATION'
   `;
 
-  const records = await queryClickhouse<Record<string, unknown>>({
+  const records = queryClickhouseStream<Record<string, unknown>>({
     query,
     params: {
       projectId,
@@ -1215,32 +1288,34 @@ export const getGenerationsForPostHog = async (
   });
 
   const baseUrl = env.NEXTAUTH_URL?.replace("/api/auth", "");
-  return records.map((record) => ({
-    timestamp: record.start_time,
-    langfuse_generation_name: record.name,
-    langfuse_trace_name: record.trace_name,
-    langfuse_url: `${baseUrl}/project/${projectId}/traces/${encodeURIComponent(record.trace_id as string)}?observation=${encodeURIComponent(record.id as string)}`,
-    langfuse_id: record.id,
-    langfuse_cost_usd: record.total_cost,
-    langfuse_input_units: record.input_tokens,
-    langfuse_output_units: record.output_tokens,
-    langfuse_total_units: record.total_tokens,
-    langfuse_session_id: record.trace_session_id,
-    langfuse_project_id: projectId,
-    langfuse_user_id: record.trace_user_id || "langfuse_unknown_user",
-    langfuse_latency: record.latency,
-    langfuse_time_to_first_token: record.time_to_first_token,
-    langfuse_release: record.trace_release,
-    langfuse_version: record.version,
-    langfuse_model: record.model,
-    langfuse_level: record.level,
-    langfuse_tags: record.trace_tags,
-    langfuse_event_version: "1.0.0",
-    $session_id: record.posthog_session_id ?? null,
-    $set: {
-      langfuse_user_url: record.user_id
-        ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
-        : null,
-    },
-  }));
+  for await (const record of records) {
+    yield {
+      timestamp: record.start_time,
+      langfuse_generation_name: record.name,
+      langfuse_trace_name: record.trace_name,
+      langfuse_url: `${baseUrl}/project/${projectId}/traces/${encodeURIComponent(record.trace_id as string)}?observation=${encodeURIComponent(record.id as string)}`,
+      langfuse_id: record.id,
+      langfuse_cost_usd: record.total_cost,
+      langfuse_input_units: record.input_tokens,
+      langfuse_output_units: record.output_tokens,
+      langfuse_total_units: record.total_tokens,
+      langfuse_session_id: record.trace_session_id,
+      langfuse_project_id: projectId,
+      langfuse_user_id: record.trace_user_id || "langfuse_unknown_user",
+      langfuse_latency: record.latency,
+      langfuse_time_to_first_token: record.time_to_first_token,
+      langfuse_release: record.trace_release,
+      langfuse_version: record.version,
+      langfuse_model: record.model,
+      langfuse_level: record.level,
+      langfuse_tags: record.trace_tags,
+      langfuse_event_version: "1.0.0",
+      $session_id: record.posthog_session_id ?? null,
+      $set: {
+        langfuse_user_url: record.user_id
+          ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
+          : null,
+      },
+    };
+  }
 };

packages/shared/src/server/repositories/observations_converters.ts

@@ -9,7 +9,6 @@ import Decimal from "decimal.js";
 import { parseClickhouseUTCDateTimeFormat } from "./clickhouse";
 import { ObservationRecordReadType } from "./definitions";
 import { parseJsonPrioritised } from "../../utils/json";
-import { jsonSchema } from "../../utils/zod";
 
 export const convertObservationToView = (
   record: ObservationRecordReadType,
@@ -62,15 +61,22 @@ export const convertObservation = (
       ? parseClickhouseUTCDateTimeFormat(record.end_time)
       : null,
     name: record.name ?? null,
-    metadata: record.metadata,
+    metadata:
+      record.metadata &&
+      Object.fromEntries(
+        Object.entries(record.metadata ?? {}).map(([key, val]) => [
+          key,
+          val && parseJsonPrioritised(val),
+        ]),
+      ),
     level: record.level as ObservationLevel,
     statusMessage: record.status_message ?? null,
     version: record.version ?? null,
     input: (record.input
-      ? jsonSchema.parse(parseJsonPrioritised(record.input))
+      ? parseJsonPrioritised(record.input)
       : null) as Prisma.JsonValue | null,
     output: (record.output
-      ? jsonSchema.parse(parseJsonPrioritised(record.output))
+      ? parseJsonPrioritised(record.output)
       : null) as Prisma.JsonValue | null,
     modelParameters: record.model_parameters
       ? JSON.parse(record.model_parameters)
@@ -126,10 +132,11 @@ export const convertObservation = (
         parseClickhouseUTCDateTimeFormat(record.start_time).getTime()
       : null,
     timeToFirstToken: record.completion_start_time
-      ? parseClickhouseUTCDateTimeFormat(
+      ? (parseClickhouseUTCDateTimeFormat(
           record.completion_start_time,
         ).getTime() -
-        parseClickhouseUTCDateTimeFormat(record.start_time).getTime()
+          parseClickhouseUTCDateTimeFormat(record.start_time).getTime()) /
+        1000
       : null,
   };
 };

packages/shared/src/server/repositories/scores.ts

@@ -3,6 +3,7 @@ import {
   commandClickhouse,
   parseClickhouseUTCDateTimeFormat,
   queryClickhouse,
+  queryClickhouseStream,
   upsertClickhouse,
 } from "./clickhouse";
 import { FilterList } from "../queries/clickhouse-sql/clickhouse-filter";
@@ -135,13 +136,16 @@ export const upsertScore = async (score: Partial<ScoreRecordReadType>) => {
   });
 };
 
-export const getScoresForTraces = async (
-  projectId: string,
-  traceIds: string[],
-  timestamp?: Date,
-  limit?: number,
-  offset?: number,
-) => {
+export type GetScoresForTracesProps = {
+  projectId: string;
+  traceIds: string[];
+  timestamp?: Date;
+  limit?: number;
+  offset?: number;
+};
+
+export const getScoresForTraces = async (props: GetScoresForTracesProps) => {
+  const { projectId, traceIds, timestamp, limit, offset } = props;
   const query = `
       select 
         *
@@ -200,20 +204,24 @@ export const getScoresForObservations = async (
   return rows.map(convertToScore);
 };
 
-export const getScoresGroupedByNameSourceType = async (projectId: string) => {
+export const getScoresGroupedByNameSourceType = async (
+  projectId: string,
+  timestamp: Date | undefined,
+) => {
   // We mainly use queries like this to retrieve filter options.
   // Therefore, we can skip final as some inaccuracy in count is acceptable.
   const query = `
-      select 
-        name,
-        source,
-        data_type
-      from scores s
-      WHERE s.project_id = {projectId: String}
-      GROUP BY name, source, data_type
-      ORDER BY count() desc
-      LIMIT 1000;
-    `;
+    select 
+      name,
+      source,
+      data_type
+    from scores s
+    WHERE s.project_id = {projectId: String}
+    ${timestamp ? `AND s.timestamp >= {timestamp: DateTime64(3)}` : ""}
+    GROUP BY name, source, data_type
+    ORDER BY count() desc
+    LIMIT 1000;
+  `;
 
   const rows = await queryClickhouse<{
     name: string;
@@ -223,6 +231,9 @@ export const getScoresGroupedByNameSourceType = async (projectId: string) => {
     query: query,
     params: {
       projectId: projectId,
+      ...(timestamp
+        ? { timestamp: convertDateToClickhouseDateTime(timestamp) }
+        : {}),
     },
   });
 
@@ -525,6 +536,27 @@ export const deleteScoresByProjectId = async (projectId: string) => {
   });
 };
 
+export const deleteScoresOlderThanDays = async (
+  projectId: string,
+  days: number,
+) => {
+  const query = `
+    DELETE FROM scores
+    WHERE project_id = {projectId: String}
+    AND timestamp < now() - INTERVAL {numDays: Int} DAYS;
+  `;
+  await commandClickhouse({
+    query: query,
+    params: {
+      projectId,
+      numDays: days,
+    },
+    clickhouseConfigs: {
+      request_timeout: 120_000, // 2 minutes
+    },
+  });
+};
+
 export const getNumericScoreHistogram = async (
   projectId: string,
   filter: FilterState,
@@ -666,11 +698,11 @@ export const getDistinctScoreNames = async (
   return rows.map((row) => row.name);
 };
 
-export const getScoresForPostHog = async (
+export const getScoresForPostHog = async function* (
   projectId: string,
   minTimestamp: Date,
   maxTimestamp: Date,
-) => {
+) {
   const query = `
     SELECT
       s.id as id,
@@ -694,7 +726,7 @@ export const getScoresForPostHog = async (
     AND t.timestamp <= {maxTimestamp: DateTime64(3)}
   `;
 
-  const records = await queryClickhouse<Record<string, unknown>>({
+  const records = queryClickhouseStream<Record<string, unknown>>({
     query,
     params: {
       projectId,
@@ -704,24 +736,26 @@ export const getScoresForPostHog = async (
   });
 
   const baseUrl = env.NEXTAUTH_URL?.replace("/api/auth", "");
-  return records.map((record) => ({
-    timestamp: record.timestamp,
-    langfuse_score_name: record.name,
-    langfuse_score_value: record.value,
-    langfuse_score_comment: record.comment,
-    langfuse_trace_name: record.trace_name,
-    langfuse_id: record.id,
-    langfuse_session_id: record.trace_session_id,
-    langfuse_project_id: projectId,
-    langfuse_user_id: record.trace_user_id || "langfuse_unknown_user",
-    langfuse_release: record.trace_release,
-    langfuse_tags: record.trace_tags,
-    langfuse_event_version: "1.0.0",
-    $session_id: record.posthog_session_id ?? null,
-    $set: {
-      langfuse_user_url: record.user_id
-        ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
-        : null,
-    },
-  }));
+  for await (const record of records) {
+    yield {
+      timestamp: record.timestamp,
+      langfuse_score_name: record.name,
+      langfuse_score_value: record.value,
+      langfuse_score_comment: record.comment,
+      langfuse_trace_name: record.trace_name,
+      langfuse_id: record.id,
+      langfuse_session_id: record.trace_session_id,
+      langfuse_project_id: projectId,
+      langfuse_user_id: record.trace_user_id || "langfuse_unknown_user",
+      langfuse_release: record.trace_release,
+      langfuse_tags: record.trace_tags,
+      langfuse_event_version: "1.0.0",
+      $session_id: record.posthog_session_id ?? null,
+      $set: {
+        langfuse_user_url: record.user_id
+          ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
+          : null,
+      },
+    };
+  }
 };

packages/shared/src/server/repositories/traces.ts

@@ -2,6 +2,7 @@ import {
   commandClickhouse,
   parseClickhouseUTCDateTimeFormat,
   queryClickhouse,
+  queryClickhouseStream,
   upsertClickhouse,
 } from "./clickhouse";
 import {
@@ -16,10 +17,7 @@ import {
 } from "../queries/clickhouse-sql/clickhouse-filter";
 import { TraceRecordReadType } from "./definitions";
 import { tracesTableUiColumnDefinitions } from "../../tableDefinitions/mapTracesTable";
-import { OrderByState } from "../../interfaces/orderBy";
-import { orderByToClickhouseSql } from "../queries/clickhouse-sql/orderby-factory";
 import { UiColumnMapping } from "../../tableDefinitions";
-import { sessionCols } from "../../tableDefinitions/mapSessionTable";
 import { convertDateToClickhouseDateTime } from "../clickhouse/client";
 import { convertClickhouseToDomain } from "./traces_converters";
 import { clickhouseSearchCondition } from "../queries/clickhouse-sql/search";
@@ -32,7 +30,7 @@ import { env } from "../../env";
 export const checkTraceExists = async (
   projectId: string,
   traceId: string,
-  timestamp: Date | undefined,
+  timestamp: Date,
   filter: FilterState,
 ): Promise<boolean> => {
   const { tracesFilter } = getProjectIdDefaultFilter(projectId, {
@@ -63,18 +61,22 @@ export const checkTraceExists = async (
   const query = `
     WITH observations_agg AS (
         SELECT
-          
             multiIf(
               arrayExists(x -> x = 'ERROR', groupArray(level)), 'ERROR',
               arrayExists(x -> x = 'WARNING', groupArray(level)), 'WARNING',
               arrayExists(x -> x = 'DEFAULT', groupArray(level)), 'DEFAULT',
               'DEBUG'
-            ) AS level,
+            ) AS aggregated_level,
+            countIf(level = 'ERROR') as error_count,
+            countIf(level = 'WARNING') as warning_count,
+            countIf(level = 'DEFAULT') as default_count,
+            countIf(level = 'DEBUG') as debug_count,
             trace_id,
             project_id
         FROM observations o FINAL 
         WHERE o.project_id = {projectId: String}
         ${timeStampFilter ? `AND o.start_time >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
+        AND o.start_time >= {timestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}
         GROUP BY trace_id, project_id
       )
     SELECT 
@@ -84,7 +86,7 @@ export const checkTraceExists = async (
     ${observationFilterRes ? `INNER JOIN observations_agg o ON t.id = o.trace_id AND t.project_id = o.project_id` : ""}
     WHERE ${tracesFilterRes.query}
     AND t.project_id = {projectId: String}
-    ${timestamp ? `AND timestamp >= {timestamp: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}` : ""}
+    AND timestamp >= {timestamp: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}
     GROUP BY t.id, t.project_id
   `;
 
@@ -267,7 +269,7 @@ export const getTracesGroupedByName = async (
       select 
         name as name,
         count(*) as count
-      from traces t
+      from traces t FINAL
       WHERE t.project_id = {projectId: String}
       AND t.name IS NOT NULL
       ${timestampFilterRes?.query ? `AND ${timestampFilterRes.query}` : ""}
@@ -383,196 +385,6 @@ export const getTracesGroupedByTags = async (props: GroupedTracesQueryProp) => {
   return rows;
 };
 
-export type SessionDataReturnType = {
-  session_id: string;
-  max_timestamp: string;
-  min_timestamp: string;
-  trace_ids: string[];
-  user_ids: string[];
-  trace_count: number;
-  trace_tags: string[];
-  total_observations: number;
-  duration: number;
-  session_usage_details: Record<string, number>;
-  session_cost_details: Record<string, number>;
-  session_input_cost: string;
-  session_output_cost: string;
-  session_total_cost: string;
-  session_input_usage: string;
-  session_output_usage: string;
-  session_total_usage: string;
-};
-
-export const getSessionsTableCount = async (props: {
-  projectId: string;
-  filter: FilterState;
-  orderBy?: OrderByState;
-  limit?: number;
-  page?: number;
-}) => {
-  const rows = await getSessionsTableGeneric<{ count: string }>({
-    select: `
-      count(session_id) as count
-    `,
-    projectId: props.projectId,
-    filter: props.filter,
-    orderBy: props.orderBy,
-    limit: props.limit,
-    page: props.page,
-  });
-
-  return rows.length > 0 ? Number(rows[0].count) : 0;
-};
-
-export const getSessionsTable = async (props: {
-  projectId: string;
-  filter: FilterState;
-  orderBy?: OrderByState;
-  limit?: number;
-  page?: number;
-}) => {
-  const rows = await getSessionsTableGeneric<SessionDataReturnType>({
-    select: `
-    session_id, 
-    max_timestamp, 
-    min_timestamp, 
-    trace_ids, 
-    user_ids, 
-    trace_count, 
-    trace_tags,
-    total_observations,
-    duration,
-    session_usage_details,
-    session_cost_details,
-    session_input_cost,
-    session_output_cost,
-    session_total_cost,
-    session_input_usage,
-    session_output_usage,
-    session_total_usage
-    `,
-    projectId: props.projectId,
-    filter: props.filter,
-    orderBy: props.orderBy,
-    limit: props.limit,
-    page: props.page,
-  });
-
-  return rows;
-};
-
-export type FetchSessionsTableProps = {
-  select: string;
-  projectId: string;
-  filter: FilterState;
-  searchQuery?: string;
-  orderBy?: OrderByState;
-  limit?: number;
-  page?: number;
-};
-
-const getSessionsTableGeneric = async <T>(props: FetchSessionsTableProps) => {
-  const { select, projectId, filter, orderBy, limit, page } = props;
-
-  const { tracesFilter, scoresFilter, observationsFilter } =
-    getProjectIdDefaultFilter(projectId, { tracesPrefix: "s" });
-
-  tracesFilter.push(...createFilterFromFilterState(filter, sessionCols));
-
-  const tracesFilterRes = tracesFilter.apply();
-  const scoresAvgFilterRes = scoresFilter.apply();
-  const observationsStatsRes = observationsFilter.apply();
-
-  const traceTimestampFilter: DateTimeFilter | undefined = tracesFilter.find(
-    (f) =>
-      f.field === "min_timestamp" &&
-      (f.operator === ">=" || f.operator === ">"),
-  ) as DateTimeFilter | undefined;
-
-  const singleTraceFilter = traceTimestampFilter
-    ? new FilterList([
-        new DateTimeFilter({
-          clickhouseTable: "traces",
-          field: "timestamp",
-          operator: traceTimestampFilter.operator,
-          value: traceTimestampFilter.value,
-        }),
-      ]).apply()
-    : undefined;
-
-  const query = `
-      WITH observations_agg AS (
-        SELECT o.trace_id,
-              count(*) as obs_count,
-              min(o.start_time) as min_start_time,
-              max(o.end_time) as max_end_time,
-              sumMap(usage_details) as sum_usage_details,
-              sumMap(cost_details) as sum_cost_details,
-              anyLast(project_id) as project_id
-        FROM observations o FINAL
-        WHERE o.project_id = {projectId: String}
-        ${traceTimestampFilter ? `AND o.start_time >= {observationsStartTime: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}` : ""}
-        GROUP BY o.trace_id
-    ),
-    session_data AS (
-        SELECT
-            t.session_id,
-            anyLast(t.project_id) as project_id,
-            max(t.timestamp) as max_timestamp,
-            min(t.timestamp) as min_timestamp,
-            groupArray(t.id) AS trace_ids,
-            groupUniqArray(t.user_id) AS user_ids,
-            count(*) as trace_count,
-            groupUniqArrayArray(t.tags) as trace_tags,
-            -- Aggregate observations data at session level
-            sum(o.obs_count) as total_observations,
-            date_diff('milliseconds', min(min_start_time), max(max_end_time)) as duration,
-            sumMap(o.sum_usage_details) as session_usage_details,
-            sumMap(o.sum_cost_details) as session_cost_details,
-            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, sumMap(o.sum_cost_details)))) as session_input_cost,
-            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, sumMap(o.sum_cost_details)))) as session_output_cost,
-            sumMap(o.sum_cost_details)['total'] as session_total_cost,          
-            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, sumMap(o.sum_usage_details)))) as session_input_usage,
-            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, sumMap(o.sum_usage_details)))) as session_output_usage,
-            sumMap(o.sum_usage_details)['total'] as session_total_usage
-        FROM traces t FINAL
-        LEFT JOIN observations_agg o
-        ON t.id = o.trace_id AND t.project_id = o.project_id
-        WHERE t.session_id IS NOT NULL
-            AND t.project_id = {projectId: String}
-            ${singleTraceFilter?.query ? ` AND ${singleTraceFilter.query}` : ""}
-        GROUP BY t.session_id
-    )
-    SELECT ${select}
-    FROM session_data s
-    WHERE ${tracesFilterRes.query ? tracesFilterRes.query : ""}
-    ${orderByToClickhouseSql(orderBy ?? null, sessionCols)}
-    ${limit !== undefined && page !== undefined ? `LIMIT {limit: Int32} OFFSET {offset: Int32}` : ""}
-    `;
-
-  const obsStartTimeValue = traceTimestampFilter
-    ? convertDateToClickhouseDateTime(traceTimestampFilter.value)
-    : null;
-
-  const res = await queryClickhouse<T>({
-    query: query,
-    params: {
-      projectId,
-      limit: limit,
-      offset: limit && page ? limit * page : 0,
-      ...tracesFilterRes.params,
-      ...observationsStatsRes.params,
-      ...scoresAvgFilterRes.params,
-      ...singleTraceFilter?.params,
-      ...(obsStartTimeValue
-        ? { observationsStartTime: obsStartTimeValue }
-        : {}),
-    },
-  });
-
-  return res;
-};
-
 export const getTracesIdentifierForSession = async (
   projectId: string,
   sessionId: string,
@@ -630,6 +442,27 @@ export const deleteTraces = async (projectId: string, traceIds: string[]) => {
   });
 };
 
+export const deleteTracesOlderThanDays = async (
+  projectId: string,
+  days: number,
+) => {
+  const query = `
+    DELETE FROM traces
+    WHERE project_id = {projectId: String}
+    AND timestamp < now() - INTERVAL {numDays: Int} DAYS;
+  `;
+  await commandClickhouse({
+    query: query,
+    params: {
+      projectId,
+      numDays: days,
+    },
+    clickhouseConfigs: {
+      request_timeout: 120_000, // 2 minutes
+    },
+  });
+};
+
 export const deleteTracesByProjectId = async (projectId: string) => {
   const query = `
     DELETE FROM traces
@@ -680,10 +513,25 @@ export const getTotalUserCount = async (
   });
 };
 
-export const getUserMetrics = async (projectId: string, userIds: string[]) => {
+export const getUserMetrics = async (
+  projectId: string,
+  userIds: string[],
+  filter: FilterState,
+) => {
   if (userIds.length === 0) {
     return [];
   }
+
+  // filter state contains date range filter for traces so far.
+  const chFilter = new FilterList(
+    createFilterFromFilterState(filter, tracesTableUiColumnDefinitions),
+  );
+  const chFilterRes = chFilter.apply();
+
+  const timestampFilter = chFilter.find(
+    (f) => f.field === "timestamp" && f.operator === ">=",
+  );
+
   // this query uses window functions on observations + traces to always get only the first row and thereby remove deduplicates
   // we filter wherever possible by project id and user id
   const query = `
@@ -713,6 +561,7 @@ export const getUserMetrics = async (projectId: string, userIds: string[]) => {
                     observations o
                 WHERE
                     o.project_id = {projectId: String }
+                    ${timestampFilter ? `AND o.start_time >= {traceTimestamp: DateTime64(3)} - ${OBSERVATIONS_TO_TRACE_INTERVAL}` : ""}
                     AND o.trace_id in (
                         SELECT
                             distinct id
@@ -721,6 +570,7 @@ export const getUserMetrics = async (projectId: string, userIds: string[]) => {
                         where
                             user_id IN ({userIds: Array(String) })
                             AND project_id = {projectId: String }
+                            ${filter.length > 0 ? `AND ${chFilterRes.query}` : ""}
                     )
                     AND o.type = 'GENERATION'
             ) as o
@@ -740,6 +590,7 @@ export const getUserMetrics = async (projectId: string, userIds: string[]) => {
                 WHERE
                     t.user_id IN ({userIds: Array(String) })
                     AND t.project_id = {projectId: String }
+                    ${filter.length > 0 ? `AND ${chFilterRes.query}` : ""}
             ) as t on t.id = o.trace_id
             and t.project_id = o.project_id
         WHERE
@@ -778,8 +629,17 @@ export const getUserMetrics = async (projectId: string, userIds: string[]) => {
     params: {
       projectId,
       userIds,
+      ...chFilterRes.params,
+      ...(timestampFilter
+        ? {
+            traceTimestamp: convertDateToClickhouseDateTime(
+              (timestampFilter as DateTimeFilter).value,
+            ),
+          }
+        : {}),
     },
   });
+
   return rows.map((row) => ({
     userId: row.user_id,
     maxTimestamp: parseClickhouseUTCDateTimeFormat(row.max_timestamp),
@@ -793,18 +653,18 @@ export const getUserMetrics = async (projectId: string, userIds: string[]) => {
   }));
 };
 
-export const getTracesForPostHog = async (
+export const getTracesForPostHog = async function* (
   projectId: string,
   minTimestamp: Date,
   maxTimestamp: Date,
-) => {
+) {
   const query = `
     WITH observations_agg AS (
       SELECT o.project_id,
              o.trace_id,
              sum(total_cost) as total_cost,
              count(*) as observation_count,
-             date_diff('milliseconds', least(min(start_time), min(end_time)), greatest(max(start_time), max(end_time))) as latency_milliseconds
+             date_diff('millisecond', least(min(start_time), min(end_time)), greatest(max(start_time), max(end_time))) as latency_milliseconds
       FROM observations o FINAL
       WHERE o.project_id = {projectId: String}
       AND o.start_time >= {minTimestamp: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}
@@ -831,7 +691,7 @@ export const getTracesForPostHog = async (
     AND t.timestamp <= {maxTimestamp: DateTime64(3)}
   `;
 
-  const records = await queryClickhouse<Record<string, unknown>>({
+  const records = queryClickhouseStream<Record<string, unknown>>({
     query,
     params: {
       projectId,
@@ -841,28 +701,30 @@ export const getTracesForPostHog = async (
   });
 
   const baseUrl = env.NEXTAUTH_URL?.replace("/api/auth", "");
-  return records.map((record) => ({
-    timestamp: record.timestamp,
-    langfuse_id: record.id,
-    langfuse_trace_name: record.name,
-    langfuse_url: `${baseUrl}/project/${projectId}/traces/${encodeURIComponent(record.id as string)}`,
-    langfuse_cost_usd: record.total_cost,
-    langfuse_count_observations: record.observation_count,
-    langfuse_session_id: record.session_id,
-    langfuse_project_id: projectId,
-    langfuse_user_id: record.user_id || "langfuse_unknown_user",
-    langfuse_latency: record.latency,
-    langfuse_release: record.release,
-    langfuse_version: record.version,
-    langfuse_tags: record.tags,
-    langfuse_event_version: "1.0.0",
-    $session_id: record.posthog_session_id ?? null,
-    $set: {
-      langfuse_user_url: record.user_id
-        ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
-        : null,
-    },
-  }));
+  for await (const record of records) {
+    yield {
+      timestamp: record.timestamp,
+      langfuse_id: record.id,
+      langfuse_trace_name: record.name,
+      langfuse_url: `${baseUrl}/project/${projectId}/traces/${encodeURIComponent(record.id as string)}`,
+      langfuse_cost_usd: record.total_cost,
+      langfuse_count_observations: record.observation_count,
+      langfuse_session_id: record.session_id,
+      langfuse_project_id: projectId,
+      langfuse_user_id: record.user_id || "langfuse_unknown_user",
+      langfuse_latency: record.latency,
+      langfuse_release: record.release,
+      langfuse_version: record.version,
+      langfuse_tags: record.tags,
+      langfuse_event_version: "1.0.0",
+      $session_id: record.posthog_session_id ?? null,
+      $set: {
+        langfuse_user_url: record.user_id
+          ? `${baseUrl}/project/${projectId}/users/${encodeURIComponent(record.user_id as string)}`
+          : null,
+      },
+    };
+  }
 };
 
 export const getTracesByIdsForAnyProject = async (traceIds: string[]) => {

packages/shared/src/server/repositories/traces_converters.ts

@@ -1,7 +1,8 @@
-import { Trace } from "@prisma/client";
+import { Prisma, Trace } from "@prisma/client";
 import { parseClickhouseUTCDateTimeFormat } from "./clickhouse";
 import { TraceRecordReadType } from "./definitions";
 import { convertDateToClickhouseDateTime } from "../clickhouse/client";
+import { parseJsonPrioritised } from "../../utils/json";
 
 export const convertTraceDomainToClickhouse = (
   trace: Trace,
@@ -43,9 +44,20 @@ export const convertClickhouseToDomain = (
     userId: record.user_id ?? null,
     sessionId: record.session_id ?? null,
     public: record.public,
-    input: record.input ?? null,
-    output: record.output ?? null,
-    metadata: record.metadata,
+    input: (record.input
+      ? parseJsonPrioritised(record.input)
+      : null) as Prisma.JsonValue | null,
+    output: (record.output
+      ? parseJsonPrioritised(record.output)
+      : null) as Prisma.JsonValue | null,
+    metadata:
+      record.metadata &&
+      Object.fromEntries(
+        Object.entries(record.metadata ?? {}).map(([key, val]) => [
+          key,
+          val && parseJsonPrioritised(val),
+        ]),
+      ),
     createdAt: parseClickhouseUTCDateTimeFormat(record.created_at),
     updatedAt: parseClickhouseUTCDateTimeFormat(record.updated_at),
     externalId: null,

packages/shared/src/server/services/StorageService.ts

@@ -1,5 +1,6 @@
 import { Readable } from "stream";
 import {
+  DeleteObjectsCommand,
   GetObjectCommand,
   ListObjectsV2Command,
   PutObjectCommand,
@@ -30,7 +31,7 @@ export interface StorageService {
 
   download(path: string): Promise<string>;
 
-  listFiles(prefix: string): Promise<string[]>;
+  listFiles(prefix: string): Promise<{ file: string; createdAt: Date }[]>;
 
   getSignedUrl(
     fileName: string,
@@ -45,6 +46,8 @@ export interface StorageService {
     contentType: string;
     contentLength: number;
   }): Promise<string>;
+
+  deleteFiles(paths: string[]): Promise<void>;
 }
 
 export class StorageServiceFactory {
@@ -201,7 +204,28 @@ class AzureBlobStorageService implements StorageService {
     }
   }
 
-  public async listFiles(prefix: string): Promise<string[]> {
+  public async deleteFiles(paths: string[]): Promise<void> {
+    try {
+      await this.createContainerIfNotExists();
+
+      await Promise.all(
+        paths.map(async (path) => {
+          const blobClient = this.client.getBlobClient(path);
+          await blobClient.deleteIfExists();
+        }),
+      );
+    } catch (err) {
+      logger.error(
+        `Failed to delete files from Azure Blob Storage ${paths}`,
+        err,
+      );
+      throw Error("Failed to delete files from Azure Blob Storage");
+    }
+  }
+
+  public async listFiles(
+    prefix: string,
+  ): Promise<{ file: string; createdAt: Date }[]> {
     try {
       await this.createContainerIfNotExists();
 
@@ -209,7 +233,10 @@ class AzureBlobStorageService implements StorageService {
       const files = [];
       for await (const blob of result) {
         if (blob.name.startsWith(prefix)) {
-          files.push(blob.name);
+          files.push({
+            file: blob.name,
+            createdAt: blob?.properties?.createdOn ?? new Date(),
+          });
         }
       }
       return files;
@@ -364,7 +391,9 @@ class S3StorageService implements StorageService {
     }
   }
 
-  public async listFiles(prefix: string): Promise<string[]> {
+  public async listFiles(
+    prefix: string,
+  ): Promise<{ file: string; createdAt: Date }[]> {
     const listCommand = new ListObjectsV2Command({
       Bucket: this.bucketName,
       Prefix: prefix,
@@ -373,7 +402,11 @@ class S3StorageService implements StorageService {
     try {
       const response = await this.client.send(listCommand);
       return (
-        response.Contents?.flatMap((file) => (file.Key ? [file.Key] : [])) ?? []
+        response.Contents?.flatMap((file) =>
+          file.Key
+            ? [{ file: file.Key, createdAt: file.LastModified ?? new Date() }]
+            : [],
+        ) ?? []
       );
     } catch (err) {
       logger.error(`Failed to list files from S3 ${prefix}`, err);
@@ -404,6 +437,37 @@ class S3StorageService implements StorageService {
     }
   }
 
+  public async deleteFiles(paths: string[]): Promise<void> {
+    const chunkSize = 900;
+    const chunks = [];
+
+    for (let i = 0; i < paths.length; i += chunkSize) {
+      chunks.push(paths.slice(i, i + chunkSize));
+    }
+
+    try {
+      for (const chunk of chunks) {
+        const command = new DeleteObjectsCommand({
+          Bucket: this.bucketName,
+          Delete: {
+            Objects: chunk.map((path) => ({ Key: path })),
+            Quiet: true,
+          },
+        });
+        const result = await this.client.send(command);
+        if (result?.Errors && result?.Errors?.length > 0) {
+          logger.error("Failed to delete files from S3", {
+            errors: result.Errors,
+          });
+          throw new Error("Failed to delete files from S3");
+        }
+      }
+    } catch (err) {
+      logger.error(`Failed to delete files from S3`, err);
+      throw new Error("Failed to delete files from S3");
+    }
+  }
+
   public async getSignedUploadUrl(params: {
     path: string;
     ttlSeconds: number;

packages/shared/src/server/services/sessions-ui-table-service.ts

@@ -0,0 +1,319 @@
+import { OrderByState } from "../../interfaces/orderBy";
+import { sessionCols } from "../../tableDefinitions/mapSessionTable";
+import { FilterState } from "../../types";
+import { convertDateToClickhouseDateTime } from "../clickhouse/client";
+import { DateTimeFilter, FilterList, orderByToClickhouseSql } from "../queries";
+import {
+  getProjectIdDefaultFilter,
+  createFilterFromFilterState,
+} from "../queries/clickhouse-sql/factory";
+import {
+  TRACE_TO_OBSERVATIONS_INTERVAL,
+  queryClickhouse,
+} from "../repositories";
+
+export type SessionDataReturnType = {
+  session_id: string;
+  max_timestamp: string;
+  min_timestamp: string;
+  trace_ids: string[];
+  user_ids: string[];
+  trace_count: number;
+  trace_tags: string[];
+};
+
+export type SessionWithMetricsReturnType = SessionDataReturnType & {
+  total_observations: number;
+  duration: number;
+  session_usage_details: Record<string, number>;
+  session_cost_details: Record<string, number>;
+  session_input_cost: string;
+  session_output_cost: string;
+  session_total_cost: string;
+  session_input_usage: string;
+  session_output_usage: string;
+  session_total_usage: string;
+};
+
+export const getSessionsTableCount = async (props: {
+  projectId: string;
+  filter: FilterState;
+  orderBy?: OrderByState;
+  limit?: number;
+  page?: number;
+}) => {
+  const rows = await getSessionsTableGeneric<{ count: string }>({
+    select: "count",
+    projectId: props.projectId,
+    filter: props.filter,
+    orderBy: props.orderBy,
+    limit: props.limit,
+    page: props.page,
+  });
+
+  return rows.length > 0 ? Number(rows[0].count) : 0;
+};
+
+export const getSessionsTable = async (props: {
+  projectId: string;
+  filter: FilterState;
+  orderBy?: OrderByState;
+  limit?: number;
+  page?: number;
+}) => {
+  const rows = await getSessionsTableGeneric<SessionDataReturnType>({
+    select: "rows",
+    projectId: props.projectId,
+    filter: props.filter,
+    orderBy: props.orderBy,
+    limit: props.limit,
+    page: props.page,
+  });
+
+  return rows.map((row) => ({
+    ...row,
+    trace_count: Number(row.trace_count),
+  }));
+};
+
+export const getSessionsWithMetrics = async (props: {
+  projectId: string;
+  filter: FilterState;
+  orderBy?: OrderByState;
+  limit?: number;
+  page?: number;
+}) => {
+  const rows = await getSessionsTableGeneric<SessionWithMetricsReturnType>({
+    select: "metrics",
+    projectId: props.projectId,
+    filter: props.filter,
+    orderBy: props.orderBy,
+    limit: props.limit,
+    page: props.page,
+  });
+
+  return rows.map((row) => ({
+    ...row,
+    trace_count: Number(row.trace_count),
+    total_observations: Number(row.total_observations),
+  }));
+};
+
+export type FetchSessionsTableProps = {
+  select: "count" | "rows" | "metrics";
+  projectId: string;
+  filter: FilterState;
+  searchQuery?: string;
+  orderBy?: OrderByState;
+  limit?: number;
+  page?: number;
+};
+
+const getSessionsTableGeneric = async <T>(props: FetchSessionsTableProps) => {
+  const { select, projectId, filter, orderBy, limit, page } = props;
+
+  let sqlSelect: string;
+  switch (select) {
+    case "count":
+      sqlSelect = "count(session_id) as count";
+      break;
+    case "rows":
+      sqlSelect = `
+          session_id, 
+          max_timestamp, 
+          min_timestamp, 
+          trace_ids, 
+          user_ids, 
+          trace_count, 
+          trace_tags`;
+      break;
+    case "metrics":
+      sqlSelect = `
+        session_id, 
+        max_timestamp, 
+        min_timestamp, 
+        trace_ids, 
+        user_ids, 
+        trace_count, 
+        trace_tags,
+        total_observations,
+        duration,
+        session_usage_details,
+        session_cost_details,
+        session_input_cost,
+        session_output_cost,
+        session_total_cost,
+        session_input_usage,
+        session_output_usage,
+        session_total_usage`;
+      break;
+    default:
+      const exhaustiveCheckDefault: never = select;
+      throw new Error(`Unknown select type: ${select}`);
+  }
+
+  const { tracesFilter } = getProjectIdDefaultFilter(projectId, {
+    tracesPrefix: "s",
+  });
+
+  tracesFilter.push(...createFilterFromFilterState(filter, sessionCols));
+
+  const tracesFilterRes = tracesFilter.apply();
+
+  const traceTimestampFilter: DateTimeFilter | undefined = tracesFilter.find(
+    (f) =>
+      f.field === "min_timestamp" &&
+      (f.operator === ">=" || f.operator === ">"),
+  ) as DateTimeFilter | undefined;
+
+  const filters = [];
+  if (traceTimestampFilter) {
+    filters.push(
+      new DateTimeFilter({
+        clickhouseTable: "traces",
+        field: "timestamp",
+        operator: traceTimestampFilter.operator,
+        value: traceTimestampFilter.value,
+      }),
+    );
+  }
+
+  const additionalSingleTraceFilter = tracesFilter.find(
+    (f) => f.field === "bookmarked" || f.field === "session_id",
+  );
+
+  if (additionalSingleTraceFilter) {
+    filters.push(additionalSingleTraceFilter);
+  }
+
+  const singleTraceFilter =
+    filters.length > 0 ? new FilterList(filters).apply() : undefined;
+
+  const hasMetricsFilter =
+    tracesFilter.find((f) =>
+      [
+        "session_total_cost",
+        "session_input_cost",
+        "session_output_cost",
+        "duration",
+        "session_total_usage",
+        "session_output_usage",
+        "session_input_usage",
+      ].includes(f.field),
+    ) ||
+    (orderBy &&
+      [
+        "totalCost",
+        "inputCost",
+        "outputCost",
+        "sessionDuration",
+        "totalTokens",
+        "outputTokens",
+        "inputTokens",
+        "usage",
+      ].includes(orderBy?.column));
+
+  const selectMetrics = select === "metrics" || hasMetricsFilter;
+
+  // We use deduplicated traces and observations CTEs instead of final to be able to use Skip indices in Clickhouse.
+  const query = `
+    WITH deduplicated_traces AS (
+      SELECT * EXCEPT input, output, metadata
+      FROM traces t
+      WHERE t.session_id IS NOT NULL 
+        AND t.project_id = {projectId: String}
+        ${singleTraceFilter?.query ? ` AND ${singleTraceFilter.query}` : ""}
+        ORDER BY event_ts DESC
+        LIMIT 1 BY id, project_id
+    ),
+    deduplicated_observations AS (
+        SELECT * 
+        FROM observations o
+        WHERE o.project_id = {projectId: String}
+        ${traceTimestampFilter ? `AND o.start_time >= {observationsStartTime: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}` : ""}
+        AND o.trace_id IN (
+          SELECT id
+          FROM deduplicated_traces
+        )
+        ORDER BY event_ts DESC
+        LIMIT 1 BY id, project_id
+    ),
+    observations_agg AS (
+      SELECT o.trace_id,
+            count(*) as obs_count,
+            min(o.start_time) as min_start_time,
+            max(o.end_time) as max_end_time,
+            sumMap(usage_details) as sum_usage_details,
+            sumMap(cost_details) as sum_cost_details,
+            anyLast(project_id) as project_id
+      FROM deduplicated_observations o
+      WHERE o.project_id = {projectId: String}
+      ${traceTimestampFilter ? `AND o.start_time >= {observationsStartTime: DateTime64(3)} - ${TRACE_TO_OBSERVATIONS_INTERVAL}` : ""}
+      GROUP BY o.trace_id
+    ),
+    session_data AS (
+        SELECT
+            t.session_id,
+            anyLast(t.project_id) as project_id,
+            max(t.timestamp) as max_timestamp,
+            min(t.timestamp) as min_timestamp,
+            groupArray(t.id) AS trace_ids,
+            groupUniqArray(t.user_id) AS user_ids,
+            count(*) as trace_count,
+            groupUniqArrayArray(t.tags) as trace_tags
+            -- Aggregate observations data at session level
+            ${
+              selectMetrics
+                ? `
+            ,
+            sum(o.obs_count) as total_observations,
+            date_diff('millisecond', min(min_start_time), max(max_end_time)) as duration,
+            sumMap(o.sum_usage_details) as session_usage_details,
+            sumMap(o.sum_cost_details) as session_cost_details,
+            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, sumMap(o.sum_cost_details)))) as session_input_cost,
+            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, sumMap(o.sum_cost_details)))) as session_output_cost,
+            sumMap(o.sum_cost_details)['total'] as session_total_cost,          
+            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, sumMap(o.sum_usage_details)))) as session_input_usage,
+            arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, sumMap(o.sum_usage_details)))) as session_output_usage,
+            sumMap(o.sum_usage_details)['total'] as session_total_usage`
+                : ""
+            }
+        FROM deduplicated_traces t
+        ${
+          selectMetrics
+            ? `LEFT JOIN observations_agg o
+        ON t.id = o.trace_id AND t.project_id = o.project_id`
+            : ""
+        }
+        WHERE t.session_id IS NOT NULL
+            AND t.project_id = {projectId: String}
+            ${singleTraceFilter?.query ? ` AND ${singleTraceFilter.query}` : ""}
+        GROUP BY t.session_id
+    )
+    SELECT ${sqlSelect}
+    FROM session_data s
+    WHERE ${tracesFilterRes.query ? tracesFilterRes.query : ""}
+    ${orderByToClickhouseSql(orderBy ?? null, sessionCols)}
+    ${limit !== undefined && page !== undefined ? `LIMIT {limit: Int32} OFFSET {offset: Int32}` : ""}
+    `;
+
+  const obsStartTimeValue = traceTimestampFilter
+    ? convertDateToClickhouseDateTime(traceTimestampFilter.value)
+    : null;
+
+  const res = await queryClickhouse<T>({
+    query: query,
+    params: {
+      projectId,
+      limit: limit,
+      offset: limit && page ? limit * page : 0,
+      ...tracesFilterRes.params,
+      ...singleTraceFilter?.params,
+      ...(obsStartTimeValue
+        ? { observationsStartTime: obsStartTimeValue }
+        : {}),
+    },
+  });
+
+  return res;
+};

packages/shared/src/server/services/traces-ui-table-service.ts

@@ -70,6 +70,10 @@ export type TracesMetricsUiReturnType = {
   scores: ScoreAggregate;
   usageDetails: Record<string, number>;
   costDetails: Record<string, number>;
+  errorCount: bigint;
+  warningCount: bigint;
+  defaultCount: bigint;
+  debugCount: bigint;
 };
 
 export const convertToUiTableRows = (
@@ -125,6 +129,10 @@ export const convertToUITableMetrics = (
       ? new Decimal(row.cost_details.output)
       : null,
     level: row.level,
+    debugCount: BigInt(row.debug_count ?? 0),
+    warningCount: BigInt(row.warning_count ?? 0),
+    errorCount: BigInt(row.error_count ?? 0),
+    defaultCount: BigInt(row.default_count ?? 0),
   };
 };
 
@@ -138,6 +146,10 @@ export type TracesTableMetricsClickhouseReturnType = {
   usage_details: Record<string, number>;
   cost_details: Record<string, number>;
   scores_avg: Array<{ name: string; avg_value: number }>;
+  error_count: number | null;
+  warning_count: number | null;
+  default_count: number | null;
+  debug_count: number | null;
 };
 
 export type FetchTracesTableProps = {
@@ -225,7 +237,11 @@ const getTracesTableGeneric = async <T>(props: FetchTracesTableProps) => {
         os.latency_milliseconds / 1000 as latency,
         os.cost_details as cost_details,
         os.usage_details as usage_details,
-        os.level as level,
+        os.aggregated_level as level,
+        os.error_count as error_count,
+        os.warning_count as warning_count,
+        os.default_count as default_count,
+        os.debug_count as debug_count,
         os.observation_count as observation_count,
         s.scores_avg as scores_avg,
         t.public as public`;
@@ -245,7 +261,6 @@ const getTracesTableGeneric = async <T>(props: FetchTracesTableProps) => {
         t.public as public`;
       break;
     default:
-      const exhaustiveCheckDefault: never = select;
       throw new Error(`Unknown select type: ${select}`);
   }
 
@@ -322,19 +337,29 @@ const getTracesTableGeneric = async <T>(props: FetchTracesTableProps) => {
       clickhouseSelect: "toDate(t.timestamp)",
       uiTableName: "timestamp_to_date",
       uiTableId: "timestamp_to_date",
-      clickhouseTableName: "observations",
+      clickhouseTableName: "traces",
+    },
+    {
+      clickhouseSelect: "t.event_ts",
+      uiTableName: "event_ts",
+      uiTableId: "event_ts",
+      clickhouseTableName: "traces",
     },
   ];
   const chOrderBy = orderByToClickhouseSql(
     [
       defaultOrder
-        ? {
-            column: "timestamp_to_date",
-            order: orderBy.order,
-          }
+        ? [
+            {
+              column: "timestamp_to_date",
+              order: orderBy.order,
+            },
+            { column: "timestamp", order: orderBy.order },
+            { column: "event_ts", order: "DESC" as "DESC" },
+          ]
         : null,
       orderBy ?? null,
-    ],
+    ].flat(),
     orderByCols,
   );
 
@@ -342,7 +367,7 @@ const getTracesTableGeneric = async <T>(props: FetchTracesTableProps) => {
   // - we only join scores and observations if we really need them to speed up default views
   // - we use FINAL on traces only in case we not need to order by something different than time. Otherwise we cannot guarantee correct reads.
   // - we filter the observations and scores as much as possible before joining them to traces.
-  // - we order by todate(timestamp), timestamp per default and do not use FINAL.
+  // - we order by todate(timestamp), event_ts desc per default and do not use FINAL.
   //   In this case, CH is able to read the data only from the latest date from disk and filtering them in memory. No need to read all data e.g. for 1 month from disk.
 
   const query = `
@@ -351,13 +376,17 @@ const getTracesTableGeneric = async <T>(props: FetchTracesTableProps) => {
         COUNT(*) AS observation_count,
           sumMap(usage_details) as usage_details,
           SUM(total_cost) AS total_cost,
-          date_diff('milliseconds', least(min(start_time), min(end_time)), greatest(max(start_time), max(end_time))) as latency_milliseconds,
+          date_diff('millisecond', least(min(start_time), min(end_time)), greatest(max(start_time), max(end_time))) as latency_milliseconds,
+          countIf(level = 'ERROR') as error_count,
+          countIf(level = 'WARNING') as warning_count,
+          countIf(level = 'DEFAULT') as default_count,
+          countIf(level = 'DEBUG') as debug_count,
           multiIf(
             arrayExists(x -> x = 'ERROR', groupArray(level)), 'ERROR',
             arrayExists(x -> x = 'WARNING', groupArray(level)), 'WARNING',
             arrayExists(x -> x = 'DEFAULT', groupArray(level)), 'DEFAULT',
             'DEBUG'
-          ) AS level,
+          ) AS aggregated_level,
           sumMap(cost_details) as cost_details,
           trace_id,
           project_id

packages/shared/src/tableDefinitions/mapObservationsTable.ts

@@ -26,6 +26,12 @@ export const observationsTableTraceUiColumnDefinitions: UiColumnMapping[] = [
 
 export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
   ...observationsTableTraceUiColumnDefinitions,
+  {
+    uiTableName: "type",
+    uiTableId: "type",
+    clickhouseTableName: "observations",
+    clickhouseSelect: 'o."type"',
+  },
   {
     uiTableName: "ID",
     uiTableId: "id",
@@ -68,7 +74,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     uiTableId: "timeToFirstToken",
     clickhouseTableName: "observations",
     clickhouseSelect:
-      "if(isNull(completion_start_time), NULL,  date_diff('milliseconds', start_time, completion_start_time) / 1000)",
+      "if(isNull(completion_start_time), NULL,  date_diff('millisecond', start_time, completion_start_time) / 1000)",
     // If we use the default of Decimal64(12), we cannot filter for more than ~40min due to an overflow
     clickhouseTypeOverwrite: "Decimal64(3)",
   },
@@ -77,7 +83,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     uiTableId: "latency",
     clickhouseTableName: "observations",
     clickhouseSelect:
-      "if(isNull(end_time), NULL, date_diff('milliseconds', start_time, end_time) / 1000)",
+      "if(isNull(end_time), NULL, date_diff('millisecond', start_time, end_time) / 1000)",
     // If we use the default of Decimal64(12), we cannot filter for more than ~40min due to an overflow
     clickhouseTypeOverwrite: "Decimal64(3)",
   },
@@ -86,7 +92,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     uiTableId: "tokensPerSecond",
     clickhouseTableName: "observations",
     clickhouseSelect:
-      "(arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, usage_details))) / (date_diff('milliseconds', start_time, end_time) / 1000))",
+      "(arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, usage_details))) / (date_diff('millisecond', start_time, end_time) / 1000))",
   },
   {
     uiTableName: "Input Cost ($)",
@@ -127,12 +133,19 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect: 'o."provided_model_name"',
   },
+  {
+    uiTableName: "Model ID",
+    uiTableId: "modelId",
+    clickhouseTableName: "observations",
+    clickhouseSelect: 'o."internal_model_id"',
+  },
   {
     uiTableName: "Input Tokens",
     uiTableId: "inputTokens",
     clickhouseTableName: "observations",
     clickhouseSelect:
       "arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, usage_details)))",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Output Tokens",
@@ -140,6 +153,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, usage_details)))",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Total Tokens",
@@ -147,6 +161,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "if(mapExists((k, v) -> (k = 'total'), usage_details), usage_details['total'], NULL)",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Usage",
@@ -154,6 +169,7 @@ export const observationsTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "if(mapExists((k, v) -> (k = 'total'), usage_details), usage_details['total'], NULL)",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Metadata",

packages/shared/src/tableDefinitions/mapSessionTable.ts

@@ -1,7 +1,7 @@
 import { UiColumnMapping } from ".";
 
 export const sessionCols: UiColumnMapping[] = [
-  // we do not access the traces scores in clichouse. We default back to the trace timestamps.
+  // we do not access the traces scores in ClickHouse. We default back to the trace timestamps.
 
   {
     uiTableName: "⭐️",
@@ -26,6 +26,8 @@ export const sessionCols: UiColumnMapping[] = [
     uiTableId: "sessionDuration",
     clickhouseTableName: "traces",
     clickhouseSelect: "duration",
+    // If we use the default of Decimal64(12), we cannot filter for more than ~40min due to an overflow
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Count Traces",
@@ -86,6 +88,8 @@ export const sessionCols: UiColumnMapping[] = [
     uiTableId: "sessionDuration",
     clickhouseTableName: "traces",
     clickhouseSelect: "duration",
+    // If we use the default of Decimal64(12), we cannot filter for more than ~40min due to an overflow
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Traces Count",

packages/shared/src/tableDefinitions/mapTracesTable.ts

@@ -12,7 +12,7 @@ export const tracesTableUiColumnDefinitions: UiColumnMapping[] = [
     uiTableName: "Level",
     uiTableId: "level",
     clickhouseTableName: "observations",
-    clickhouseSelect: "level",
+    clickhouseSelect: "aggregated_level",
   },
   {
     uiTableName: "ID",
@@ -68,12 +68,37 @@ export const tracesTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "traces",
     clickhouseSelect: "tags",
   },
+  {
+    uiTableName: "Warning Level Count",
+    uiTableId: "warningCount",
+    clickhouseTableName: "observations",
+    clickhouseSelect: "warning_count",
+  },
+  {
+    uiTableName: "Error Level Count",
+    uiTableId: "errorCount",
+    clickhouseTableName: "observations",
+    clickhouseSelect: "error_count",
+  },
+  {
+    uiTableName: "Default Level Count",
+    uiTableId: "defaultCount",
+    clickhouseTableName: "observations",
+    clickhouseSelect: "default_count",
+  },
+  {
+    uiTableName: "Debug Level Count",
+    uiTableId: "debugCount",
+    clickhouseTableName: "observations",
+    clickhouseSelect: "debug_count",
+  },
   {
     uiTableName: "Input Tokens",
     uiTableId: "inputTokens",
     clickhouseTableName: "observations",
     clickhouseSelect:
       "arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'input') > 0, usage_details)))",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Output Tokens",
@@ -81,6 +106,7 @@ export const tracesTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "arraySum(mapValues(mapFilter(x -> positionCaseInsensitive(x.1, 'output') > 0, usage_details)))",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Total Tokens",
@@ -88,6 +114,7 @@ export const tracesTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "if(mapExists((k, v) -> (k = 'total'), usage_details), usage_details['total'], NULL)",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Usage",
@@ -95,6 +122,7 @@ export const tracesTableUiColumnDefinitions: UiColumnMapping[] = [
     clickhouseTableName: "observations",
     clickhouseSelect:
       "if(mapExists((k, v) -> (k = 'total'), usage_details), usage_details['total'], NULL)",
+    clickhouseTypeOverwrite: "Decimal64(3)",
   },
   {
     uiTableName: "Scores",

packages/shared/src/tableDefinitions/tracesTable.ts

@@ -88,6 +88,30 @@ export const tracesTableCols: ColumnDefinition[] = [
     internal: 'generation_metrics."completionTokens"',
     nullable: true,
   },
+  {
+    name: "Error Level Count",
+    id: "errorCount",
+    type: "number",
+    internal: 'generation_metrics."errorCount"',
+  },
+  {
+    name: "Warning Level Count",
+    id: "warningCount",
+    type: "number",
+    internal: 'generation_metrics."warningCount"',
+  },
+  {
+    name: "Default Level Count",
+    id: "defaultCount",
+    type: "number",
+    internal: 'generation_metrics."defaultCount"',
+  },
+  {
+    name: "Debug Level Count",
+    id: "debugCount",
+    type: "number",
+    internal: 'generation_metrics."debugCount"',
+  },
   {
     name: "Total Tokens",
     id: "totalTokens",
@@ -102,7 +126,6 @@ export const tracesTableCols: ColumnDefinition[] = [
     internal: 'generation_metrics."totalTokens"',
     nullable: true,
   },
-
   {
     name: "Scores",
     id: "scores_avg",

packages/shared/src/utils/json.ts

@@ -1,6 +1,4 @@
-import { z } from "zod";
-import { merge } from "lodash";
-import { JsonNested, jsonSchema, jsonSchemaNullable } from "./zod";
+import { JsonNested } from "./zod";
 
 /**
  * Deeply parses a JSON string or object for nested stringified JSON
@@ -10,7 +8,6 @@ import { JsonNested, jsonSchema, jsonSchemaNullable } from "./zod";
 export function deepParseJson(json: unknown): unknown {
   if (typeof json === "string") {
     try {
-      // eslint-disable-next-line @typescript-eslint/no-unsafe-assignment
       const parsed = JSON.parse(json);
       if (typeof parsed === "number") return json; // numbers that were strings in the input should remain as strings
       return deepParseJson(parsed); // Recursively parse parsed value
@@ -40,61 +37,12 @@ export function deepParseJson(json: unknown): unknown {
   return json;
 }
 
-export const mergeJson = (
-  json1?: z.infer<typeof jsonSchema>,
-  json2?: z.infer<typeof jsonSchema>,
-) => {
-  if (json1 === undefined) {
-    return json2;
-  }
-  return merge(json1, json2);
-};
-
 export const parseJsonPrioritised = (
   json: string,
-): z.infer<typeof jsonSchema> | string | undefined => {
+): JsonNested | string | undefined => {
   try {
-    const parsedJson = JSON.parse(json);
-
-    if (Object.keys(parsedJson).length === 0) {
-      return parsedJson;
-    }
-
-    const parsedArray = z.array(jsonSchemaNullable).safeParse(parsedJson);
-    if (parsedArray.success) {
-      return parsedArray.data;
-    }
-
-    const parsedObject = z.record(jsonSchemaNullable).safeParse(parsedJson);
-    if (parsedObject.success) {
-      return parsedObject.data;
-    }
-
-    return jsonSchema.parse(parsedJson);
+    return JSON.parse(json);
   } catch (error) {
-    const parsed = jsonSchema.safeParse(json);
-
-    return parsed.success ? parsed.data : json;
+    return json;
   }
 };
-
-export const convertRecordToJsonSchema = (
-  record: Record<string, string>,
-): JsonNested | undefined => {
-  const jsonSchema: JsonNested = {};
-
-  // if record is empty, return undefined
-  if (Object.keys(record).length === 0) {
-    return undefined;
-  }
-
-  for (const key in record) {
-    try {
-      jsonSchema[key] = JSON.parse(record[key]);
-    } catch (e) {
-      jsonSchema[key] = record[key];
-    }
-  }
-
-  return jsonSchema;
-};

packages/shared/src/utils/stringChecks.ts

@@ -4,22 +4,27 @@ export function getIsCharOrUnderscore(value: string): boolean {
   return charOrUnderscore.test(value);
 }
 
+// Regex for valid variable names (letters, underscores, starting with letter)
+export const VARIABLE_REGEX = /^[a-zA-Z][a-zA-Z_]*$/;
+
+// Regex to find variables in mustache syntax
+export const MUSTACHE_REGEX = /{{([^{}]*)}}+/g;
+
+// Regex to find multiline variables
+export const MULTILINE_VARIABLE_REGEX = /{{[^}]*\n[^}]*}}/g;
+
+// Regex to find unclosed variables
+export const UNCLOSED_VARIABLE_REGEX = /{{(?![^{]*}})/g;
+
+export function isValidVariableName(variable: string): boolean {
+  return VARIABLE_REGEX.test(variable);
+}
+
 export function extractVariables(mustacheString: string): string[] {
-  const mustacheRegex = /\{\{(.*?)\}\}/g;
-  const uniqueVariables = new Set<string>();
-
-  for (const match of mustacheString.matchAll(mustacheRegex)) {
-    uniqueVariables.add(match[1]);
-  }
-
-  for (const variable of uniqueVariables) {
-    // if validated fails, remove from set
-    if (!getIsCharOrUnderscore(variable)) {
-      uniqueVariables.delete(variable);
-    }
-  }
-
-  return Array.from(uniqueVariables);
+  const matches = Array.from(mustacheString.matchAll(MUSTACHE_REGEX))
+    .map((match) => match[1])
+    .filter(isValidVariableName);
+  return [...new Set(matches)];
 }
 
 export function stringifyValue(value: unknown) {

packages/shared/src/utils/zod.ts

@@ -23,18 +23,18 @@ type Json = Root | { [key: string]: JsonNested } | JsonNested[];
 // Here, you define the schema recursively
 export const jsonSchemaNullable: z.ZodType<JsonNested> = z.lazy(() =>
   z.union([
-    nestedLiteralSchema,
     z.array(jsonSchemaNullable),
     z.record(jsonSchemaNullable),
+    nestedLiteralSchema,
   ]),
 );
 
 // Root schema that does not allow nulls at the root level
 export const jsonSchema: z.ZodType<Json> = z.lazy(() =>
   z.union([
-    rootLiteralSchema,
     z.array(jsonSchemaNullable),
     z.record(jsonSchemaNullable),
+    rootLiteralSchema,
   ]),
 );