docker-compose.yml

# Dataing Self-Hosted Docker Compose
# Production-ready deployment for autonomous data quality investigation
#
# Usage:
#   docker compose up -d                    # Start all services
#   docker compose --profile observability  # Include Jaeger/OTEL
#   docker compose down -v                  # Stop and remove volumes
#
# Required environment variables:
#   ANTHROPIC_API_KEY       - Anthropic API key for LLM features
#   DATADR_ENCRYPTION_KEY   - Fernet key for encrypting datasource credentials
#
# Access points:
#   Frontend:     http://localhost:3000
#   API:          http://localhost:8000
#   Temporal UI:  http://localhost:8233

services:
  # ===========================================================================
  # Core Infrastructure
  # ===========================================================================
  postgres:
    image: pgvector/pgvector:pg16
    environment:
      POSTGRES_DB: dataing
      POSTGRES_USER: dataing
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-dataing}
    volumes:
      - dataing-pgdata:/var/lib/postgresql/data
      - ./infra/init-temporal-db.sh:/docker-entrypoint-initdb.d/init-temporal-db.sh:ro
    ports:
      - "5432:5432"
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U dataing -d dataing"]
      interval: 5s
      timeout: 5s
      retries: 5
    restart: unless-stopped

  redis:
    image: redis:7-alpine
    command: ["redis-server", "--appendonly", "yes"]
    ports:
      - "6379:6379"
    volumes:
      - dataing-redis:/data
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 5
    restart: unless-stopped

  # ===========================================================================
  # Temporal Server
  # ===========================================================================
  temporal:
    image: temporalio/auto-setup:latest
    environment:
      DB: postgres12
      DB_PORT: 5432
      POSTGRES_SEEDS: postgres
      POSTGRES_USER: temporal
      POSTGRES_PWD: temporal  # pragma: allowlist secret
      SKIP_DB_CREATE: "true"
    ports:
      - "7233:7233"
    depends_on:
      postgres:
        condition: service_healthy
    healthcheck:
      test: ["CMD-SHELL", "tctl --address $(hostname -i):7233 namespace list | grep -q temporal-system"]
      interval: 10s
      timeout: 5s
      retries: 10
      start_period: 60s
    restart: unless-stopped

  temporal-ui:
    image: temporalio/ui:latest
    environment:
      TEMPORAL_ADDRESS: temporal:7233
      TEMPORAL_CORS_ORIGINS: http://localhost:3000
    ports:
      - "8233:8080"
    depends_on:
      temporal:
        condition: service_healthy
    healthcheck:
      test: ["CMD", "wget", "-q", "--spider", "http://127.0.0.1:8080/"]
      interval: 10s
      timeout: 5s
      retries: 3
    restart: unless-stopped

  # ===========================================================================
  # Database Migrations (runs once on startup)
  # ===========================================================================
  db-migrate:
    image: postgres:16-alpine
    volumes:
      - ./python-packages/dataing/migrations:/migrations:ro
      - ./infra/init-app-db.sh:/init-app-db.sh:ro
    environment:
      PGHOST: postgres
      PGPORT: 5432
      PGUSER: dataing
      PGPASSWORD: ${POSTGRES_PASSWORD:-dataing}
      PGDATABASE: dataing
      MIGRATIONS_DIR: /migrations
    entrypoint: ["/bin/sh", "/init-app-db.sh"]
    depends_on:
      postgres:
        condition: service_healthy
    restart: "no"

  # ===========================================================================
  # Dataing Application
  # ===========================================================================
  api:
    build:
      context: .
      dockerfile: infra/Dockerfile.backend
    env_file:
      - path: .env
        required: false
    environment:
      DATABASE_URL: postgresql://dataing:${POSTGRES_PASSWORD:-dataing}@postgres:5432/dataing
      APP_DATABASE_URL: postgresql://dataing:${POSTGRES_PASSWORD:-dataing}@postgres:5432/dataing
      TEMPORAL_HOST: temporal:7233
      TEMPORAL_NAMESPACE: ${TEMPORAL_NAMESPACE:-default}
      TEMPORAL_TASK_QUEUE: ${TEMPORAL_TASK_QUEUE:-investigations}
      INVESTIGATION_ENGINE: temporal
      DATADR_ENCRYPTION_KEY: ${DATADR_ENCRYPTION_KEY:-}
      ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:-}
      REDIS_HOST: redis
      REDIS_PORT: 6379
      LLM_MODEL: ${LLM_MODEL:-claude-sonnet-4-20250514}
    ports:
      - "8000:8000"
    depends_on:
      db-migrate:
        condition: service_completed_successfully
      postgres:
        condition: service_healthy
      redis:
        condition: service_healthy
      temporal:
        condition: service_healthy
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 10s
    restart: unless-stopped

  worker:
    build:
      context: .
      dockerfile: infra/Dockerfile.backend
    command: ["python", "-m", "dataing.entrypoints.temporal_worker"]
    env_file:
      - path: .env
        required: false
    environment:
      DATABASE_URL: postgresql://dataing:${POSTGRES_PASSWORD:-dataing}@postgres:5432/dataing
      APP_DATABASE_URL: postgresql://dataing:${POSTGRES_PASSWORD:-dataing}@postgres:5432/dataing
      TEMPORAL_HOST: temporal:7233
      TEMPORAL_NAMESPACE: ${TEMPORAL_NAMESPACE:-default}
      TEMPORAL_TASK_QUEUE: ${TEMPORAL_TASK_QUEUE:-investigations}
      DATADR_ENCRYPTION_KEY: ${DATADR_ENCRYPTION_KEY:-}
      ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:-}
      REDIS_HOST: redis
      REDIS_PORT: 6379
      LLM_MODEL: ${LLM_MODEL:-claude-sonnet-4-20250514}
    depends_on:
      db-migrate:
        condition: service_completed_successfully
      postgres:
        condition: service_healthy
      redis:
        condition: service_healthy
      temporal:
        condition: service_healthy
    healthcheck:
      test: ["CMD-SHELL", "python -c \"import asyncio; from temporalio.client import Client; asyncio.run(Client.connect('temporal:7233'))\" 2>/dev/null"]
      interval: 15s
      timeout: 10s
      retries: 5
      start_period: 30s
    restart: unless-stopped

  frontend:
    build:
      context: .
      dockerfile: infra/Dockerfile.frontend
    environment:
      API_URL: http://api:8000
    ports:
      - "3000:3000"
    depends_on:
      api:
        condition: service_healthy
    healthcheck:
      test: ["CMD", "wget", "-q", "--spider", "http://127.0.0.1:3000/nginx-health"]
      interval: 10s
      timeout: 5s
      retries: 3
    restart: unless-stopped

  # ===========================================================================
  # Observability (optional, enable with --profile observability)
  # ===========================================================================
  otel-collector:
    image: otel/opentelemetry-collector-contrib:0.96.0
    profiles: ["observability"]
    command: ["--config=/etc/otelcol-contrib/config.yaml"]
    volumes:
      - ./infra/otel-collector-config.yaml:/etc/otelcol-contrib/config.yaml:ro
    ports:
      - "4317:4317"
      - "4318:4318"
    depends_on:
      - jaeger
    restart: unless-stopped

  jaeger:
    image: jaegertracing/all-in-one:1.76.0
    profiles: ["observability"]
    environment:
      COLLECTOR_OTLP_ENABLED: "true"
    ports:
      - "16686:16686"
      - "14268:14268"
    healthcheck:
      test: ["CMD", "wget", "-q", "--spider", "http://localhost:16686"]
      interval: 5s
      timeout: 3s
      retries: 5
    restart: unless-stopped

volumes:
  dataing-pgdata:
  dataing-redis: