mirror of https://github.com/hanzoai/insights.git synced 2026-11-03 00:52:32 +00:00

Files

T

zeekay 1b367e2b62 chore: rebase the tree onto upstream master

The tree was a debranded snapshot of upstream taken around 2026-02-20, and
upstream has moved about five months since. This replaces it with upstream
master, mechanically re-derived: bin/debrand does the rename, and the
pipeline and squashed migrations are restored from the shipped tree.

Not a merge. The previous import replaced the tree wholesale in a single
commit, so there is no shared content history to merge against. Reproducing
the rename on the new tree is what actually converges.

What survives: .hanzo/workflows, the Dockerfiles and bin/docker* entrypoints,
compose files, LLM.md, and insights/migrations. The squashed initial
migration is load-bearing — the live database records it by name, so
upstream's ~918 individual migrations cannot come back.

Nine products are dropped as unreferenced. Twenty-one more that we do not run
are kept anyway because the core app still imports them; unpicking those is a
refactor and does not belong in a rebase.

This base does not build yet, and the reasons are not in the rename:
upstream now wants Python 3.13.13 against our 3.12.12 and Django 5.2 against
our 4.2, and the dependency manifests still name upstream packages that we
republish under our own names. Those are decisions, not renames.

2026-08-01 15:44:40 -07:00

3.2 KiB

Raw Permalink Blame History

OpenSearch — local LLM trace reverse index

Local Docker OpenSearch cluster + checked-in llm-traces-v0_1 index template. Stage 0 of the LLMA reverse-index POC; the (forthcoming) rust/opensearch-indexer writes to localhost:9201. Prod uses managed AWS OpenSearch Service via Terraform in insights-cloud-infra.

What's running

Service	Image	Host port	Notes
`opensearch`	`opensearchproject/opensearch:2.13.0`	`127.0.0.1:9201` → `9200`	single-node, security plugin off, 512m heap, ephemeral data
`opensearch-dashboards`	`opensearchproject/opensearch-dashboards:2.13.0`	`127.0.0.1:5601` → `5601`	UI for ad-hoc queries
`opensearch-init`	`curlimages/curl:8.10.1`	—	one-shot bootstrap; exits 0 after applying template + alias

Activated by the opensearch_search capability (in ai_features intent). Run via bin/start ai_features, or directly: docker compose -f docker-compose.dev.yml -f docker-compose.profiles.yml --profile opensearch up -d.

Verify

curl -s http://localhost:9201/_cluster/health         # status: green or yellow
curl -s 'http://localhost:9201/_cat/indices/llm-traces-v0_1-*?v'    # llm-traces-v0_1-000001
curl -s http://localhost:9201/_alias/llm-traces       # is_write_index: true

Dashboards UI: http://localhost:5601 → Dev Tools → GET llm-traces/_mapping.

Reset (wipe data, re-bootstrap)

docker compose -f docker-compose.dev.yml down -v opensearch opensearch-dashboards opensearch-init
docker compose -f docker-compose.dev.yml -f docker-compose.profiles.yml --profile opensearch up -d

Reapply template after editing the JSON

docker compose -f docker-compose.dev.yml run --rm opensearch-init

Idempotent — safe to run any time. Template applies to new indices only; to pick up mapping changes on the existing backing index, do a full reset.

Notes

opensearch-init shows Exited (0) in docker ps -a — that's expected. It's a one-shot bootstrap container, not a daemon.
Port 9200 is owned by Temporal's Elasticsearch — that's why OpenSearch publishes on 9201.
No ISM locally. Hot → UltraWarm requires UltraWarm-typed nodes (S3-backed); a single-node cluster can't provide them. Local is hot-only; prod uses ISM via Terraform.
Security plugin is off. Bind is loopback-only (127.0.0.1) so the open cluster isn't reachable off-host. AWS OpenSearch Service authenticates via SigV4/IAM at the load balancer instead — no in-cluster RBAC there either.
Schema source of truth: llm-traces-v0_1.template.json. Same file is consumed by the init container locally and (eventually) by Terraform in prod.
Local cluster status is yellow, not green. Template sets number_of_replicas: 1 for prod redundancy; the single-node local cluster can't allocate the replica, so the index sits in yellow. Functionally fine for dev.

3.2 KiB Raw Permalink Blame History