dev/deploy

Hanzo Platform - Deployment Guide

February 2026: Production services run on DOKS K8s clusters (hanzo-k8s, lux-k8s). DO App Platform is decommissioned. The configs below are for local development and reference.

This directory contains deployment configurations for the Hanzo Platform.

Architecture

Hanzo Node uses embedded storage and consensus:

• sled: High-performance embedded key-value store (Rust)
• Lux consensus: Decentralized consensus via libp2p networking

No external databases (PostgreSQL, Redis, MinIO) are required. All state is persisted locally in /app/data and replicated across nodes via Lux consensus.

Directory Structure

deploy/
├── do/                     # Digital Ocean App Platform
│   └── app.yaml           # App Platform specification
├── terraform/              # Infrastructure as Code
│   ├── main.tf            # Main Terraform configuration
│   └── templates/         # Cloud-init templates
│       └── node-cloud-init.yaml
├── scripts/               # Deployment scripts
│   └── deploy.sh         # Main deployment script
├── compose.yml            # Local development stack
├── Makefile              # Make targets for deployment
└── README.md             # This file

Prerequisites

1. Digital Ocean Account with:

   • API token (doctl auth init)
   • Container registry created

2. Tools Required:

   • terraform >= 1.5.0
   • doctl (DO CLI)
   • docker and docker compose
   • make

3. GitHub Secrets (for CI/CD):

   • DIGITALOCEAN_ACCESS_TOKEN

Quick Start

Local Development

# Start the full local stack
make dev

# Or with monitoring (Prometheus + Grafana)
make dev-full

# View logs
make logs

# Check health
make health

# Stop everything
make down

# Clean up (removes volumes)
make clean

Service URLs:

• Console: http://localhost:3001
• Node API: http://localhost:8080
• Platform API: http://localhost:8000
• Grafana: http://localhost:3000 (admin/hanzo)

Deploy to Digital Ocean

1. Configure Terraform

# Copy example vars
cp terraform/terraform.tfvars.example terraform/terraform.tfvars

# Edit with your values
vim terraform/terraform.tfvars

2. Plan and Apply Infrastructure

# Initialize Terraform
make tf-init

# Plan changes
make plan ENV=staging

# Apply changes
make apply ENV=staging

3. Deploy App Platform

# Create the app
make app-create

# Or update existing
make app-update

# Trigger deployment
make app-deploy

# View logs
make app-logs

4. Full Deployment

# Build, push, and deploy everything
make deploy

Architecture Details

Services

Service	Port	Description
console	3001	Next.js admin dashboard
hanzo-node	8080, 9090, 9000	Core compute node (sled + Lux)
platform-api	8000	Compute pools API

Data Layer

Embedded sled storage:

• ACID-compliant key-value store
• Zero-copy reads for high performance
• Automatic compaction and recovery
• Data persisted to /app/data

Lux consensus:

• libp2p-based P2P networking
• State replication across nodes
• Byzantine fault tolerant
• No central coordination required

Infrastructure (Terraform)

• VPC: Private network (10.10.0.0/16)
• Droplets: Compute nodes with Docker
• Load Balancer: HTTPS termination
• Firewall: Network security rules
• DNS: Subdomain records

App Platform

DO App Platform provides:

• Automatic deployments on git push
• Managed SSL certificates
• Auto-scaling
• Built-in monitoring

Configuration

Environment Variables

Variable	Description	Default
HANZO_NODE_URL	Node API URL	http://localhost:8080
HANZO_DATA_DIR	Data directory	/app/data
HANZO_NETWORK_ID	Network ID	1337 (local)

Terraform Variables

See terraform/terraform.tfvars.example for all available options.

Key variables:

• do_token: DO API token
• environment: production/staging/development
• region: DO region (nyc3, sfo3, etc.)
• node_count: Number of compute nodes

CI/CD

GitHub Actions workflow (.github/workflows/do-deploy.yml):

1. On PR: Terraform plan, validate app spec
2. On push to main: Build images, apply Terraform, deploy app
3. Manual dispatch: Choose environment and action

Required Secrets

• DIGITALOCEAN_ACCESS_TOKEN: DO API token

Monitoring

Local Development

Start with monitoring profile:

make dev-full

Access:

• Prometheus: http://localhost:9091
• Grafana: http://localhost:3000

Production

DO provides built-in monitoring. Additional:

• Node metrics exposed at /v1/metrics
• Prometheus scrape config in hanzo-node/config/prometheus.yml
• Alerts configured in Terraform

Data Management

Backup

# Local: backup sled data
docker compose exec hanzo-node tar -czvf /tmp/backup.tar.gz /app/data
docker compose cp hanzo-node:/tmp/backup.tar.gz ./backup.tar.gz

# Production: snapshot droplet volumes
doctl compute droplet-action snapshot <droplet-id> --snapshot-name "hanzo-backup-$(date +%Y%m%d)"

Restore

# Local: restore sled data
docker compose cp ./backup.tar.gz hanzo-node:/tmp/backup.tar.gz
docker compose exec hanzo-node tar -xzvf /tmp/backup.tar.gz -C /
docker compose restart hanzo-node

Troubleshooting

Common Issues

1. Port conflicts: Check if ports 3001, 8080 are free
2. Docker memory: Ensure Docker has >= 4GB RAM
3. Terraform state: Use remote backend for team deployments
4. Data corruption: Check sled logs; restore from backup if needed

Health Checks

# Local
make health

# Production
make health-prod

Logs

# All services
make logs

# Specific service
make logs-node
make logs-console

# Production
make app-logs

Node Data

# Check sled data directory
docker compose exec hanzo-node ls -la /app/data

# View sled metrics
curl http://localhost:8080/v1/metrics | grep sled

Security Notes

1. Never commit:

   • terraform.tfvars
   • .env files with secrets
   • Private keys

2. Production hardening:

   • Restrict SSH access IPs
   • Use private networking
   • Enable backups
   • Regular security updates

3. Secrets management:

   • Use DO secrets for sensitive env vars
   • Rotate credentials regularly
   • Use least-privilege access

Support

• Documentation: https://docs.hanzo.ai
• Issues: https://github.com/hanzoai/dev/issues
• Email: [email protected]