provenance/deploy/docker-compose.yml

name: provenance

# One stack stands up the whole system. Configuration is entirely env-driven
# (see .env.example). Run from this directory: `docker compose up -d`.
#
# backend/frontend are PULLED from the public registry (git.jpaul.io); CI pushes
# them to the LAN endpoint (192.168.0.2:1234). For local building instead of
# pulling, layer the dev override:
#   docker compose -f docker-compose.yml -f docker-compose.dev.yml up -d --build

services:
  postgres:
    # pgvector image = Postgres + pgvector; pg_trgm ships in contrib.
    image: pgvector/pgvector:pg17
    environment:
      POSTGRES_USER: ${POSTGRES_USER:-provenance}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-provenance}
      POSTGRES_DB: ${POSTGRES_DB:-provenance}
    volumes:
      - pgdata:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-provenance} -d ${POSTGRES_DB:-provenance}"]
      interval: 5s
      timeout: 5s
      retries: 10
    restart: unless-stopped

  minio:
    image: minio/minio:latest
    command: server /data --console-address ":9001"
    environment:
      MINIO_ROOT_USER: ${MINIO_ROOT_USER:-provenance}
      MINIO_ROOT_PASSWORD: ${MINIO_ROOT_PASSWORD:-change-me-too}
    volumes:
      - miniodata:/data
    healthcheck:
      test: ["CMD-SHELL", "mc ready local || exit 1"]
      interval: 10s
      timeout: 5s
      retries: 10
    restart: unless-stopped

  # One-shot schema migration: runs `alembic upgrade head` and exits. Backend
  # and worker wait for it to finish, so on `docker compose up` the schema is
  # always current before the app serves traffic — no manual migrate step.
  # NOTE: a pure Watchtower image-swap recreates only the long-running
  # containers, not this one-shot job, so Watchtower deploys should be paired
  # with a `compose up` (see deploy docs) to re-run migrations.
  migrate:
    image: git.jpaul.io/justin/provenance-backend:${IMAGE_TAG:-test-main}
    command: ["uv", "run", "--no-dev", "alembic", "upgrade", "head"]
    labels:
      com.centurylinklabs.watchtower.enable: "true"
    environment:
      APP_ENV: ${APP_ENV:-development}
      DATABASE_URL: ${DATABASE_URL:-postgresql+asyncpg://provenance:provenance@postgres:5432/provenance}
    depends_on:
      postgres:
        condition: service_healthy
    restart: "no"

  backend:
    image: git.jpaul.io/justin/provenance-backend:${IMAGE_TAG:-test-main}
    labels:
      com.centurylinklabs.watchtower.enable: "true"
    environment:
      APP_ENV: ${APP_ENV:-development}
      # Self-migrate on start so a Watchtower in-place image swap applies any new
      # migrations (idempotent). The one-shot `migrate` service covers the same
      # for `compose up`; the depends_on below serializes them so they never run
      # alembic concurrently.
      RUN_MIGRATIONS: "1"
      DATABASE_URL: ${DATABASE_URL:-postgresql+asyncpg://provenance:provenance@postgres:5432/provenance}
      # Instance owner/operator — the account(s) with instance-admin rights.
      OWNER_EMAIL: ${OWNER_EMAIL:-}
      S3_ENDPOINT_URL: ${S3_ENDPOINT_URL:-http://minio:9000}
      S3_BUCKET: ${S3_BUCKET:-provenance}
      S3_ACCESS_KEY: ${S3_ACCESS_KEY:-provenance}
      S3_SECRET_KEY: ${S3_SECRET_KEY:-change-me-too}
      S3_REGION: ${S3_REGION:-us-east-1}
      # Email / mailer — verification + password-reset links. APP_BASE_URL is the
      # base for those links; MAILER=smtp activates the SMTP_* settings.
      APP_BASE_URL: ${APP_BASE_URL:-http://localhost}
      REQUIRE_EMAIL_VERIFICATION: ${REQUIRE_EMAIL_VERIFICATION:-false}
      MAILER: ${MAILER:-console}
      SMTP_HOST: ${SMTP_HOST:-}
      SMTP_PORT: ${SMTP_PORT:-587}
      SMTP_USERNAME: ${SMTP_USERNAME:-}
      SMTP_PASSWORD: ${SMTP_PASSWORD:-}
      SMTP_FROM: ${SMTP_FROM:-Provenance <no-reply@provenance.local>}
      # Model providers (AI assistant + embeddings). Each activates when its key
      # is set; DEFAULT_*_PROVIDER picks the default. 'null' keeps AI off.
      DEFAULT_LLM_PROVIDER: ${DEFAULT_LLM_PROVIDER:-null}
      DEFAULT_EMBEDDING_PROVIDER: ${DEFAULT_EMBEDDING_PROVIDER:-null}
      LLM_MAX_TOKENS: ${LLM_MAX_TOKENS:-4096}
      EMBEDDING_DIMENSIONS: ${EMBEDDING_DIMENSIONS:-1536}
      ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:-}
      ANTHROPIC_MODEL: ${ANTHROPIC_MODEL:-claude-opus-4-8}
      OPENAI_API_KEY: ${OPENAI_API_KEY:-}
      OPENAI_BASE_URL: ${OPENAI_BASE_URL:-https://api.openai.com/v1}
      OPENAI_MODEL: ${OPENAI_MODEL:-gpt-4o}
      OPENAI_EMBEDDING_MODEL: ${OPENAI_EMBEDDING_MODEL:-text-embedding-3-small}
      XAI_API_KEY: ${XAI_API_KEY:-}
      XAI_BASE_URL: ${XAI_BASE_URL:-https://api.x.ai/v1}
      XAI_MODEL: ${XAI_MODEL:-grok-2-latest}
      OLLAMA_ENABLED: ${OLLAMA_ENABLED:-false}
      OLLAMA_BASE_URL: ${OLLAMA_BASE_URL:-http://localhost:11434/v1}
      OLLAMA_MODEL: ${OLLAMA_MODEL:-llama3.1}
      OLLAMA_EMBEDDING_MODEL: ${OLLAMA_EMBEDDING_MODEL:-nomic-embed-text}
    depends_on:
      postgres:
        condition: service_healthy
      minio:
        condition: service_healthy
      migrate:
        condition: service_completed_successfully
    healthcheck:
      test:
        - CMD-SHELL
        - >-
          python -c "import urllib.request,sys;
          sys.exit(0 if urllib.request.urlopen('http://localhost:8000/health').status==200 else 1)"
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 20s
    restart: unless-stopped

  # Background worker — same image as the backend, run in worker mode.
  # First job: the scheduled soft-delete purge (and media object cleanup).
  worker:
    image: git.jpaul.io/justin/provenance-backend:${IMAGE_TAG:-test-main}
    command: ["uv", "run", "--no-dev", "python", "-m", "app.worker"]
    labels:
      com.centurylinklabs.watchtower.enable: "true"
    environment:
      APP_ENV: ${APP_ENV:-development}
      DATABASE_URL: ${DATABASE_URL:-postgresql+asyncpg://provenance:provenance@postgres:5432/provenance}
      S3_ENDPOINT_URL: ${S3_ENDPOINT_URL:-http://minio:9000}
      S3_BUCKET: ${S3_BUCKET:-provenance}
      S3_ACCESS_KEY: ${S3_ACCESS_KEY:-provenance}
      S3_SECRET_KEY: ${S3_SECRET_KEY:-change-me-too}
      S3_REGION: ${S3_REGION:-us-east-1}
    depends_on:
      postgres:
        condition: service_healthy
      minio:
        condition: service_healthy
      migrate:
        condition: service_completed_successfully
    restart: unless-stopped

  frontend:
    image: git.jpaul.io/justin/provenance-frontend:${IMAGE_TAG:-test-main}
    labels:
      com.centurylinklabs.watchtower.enable: "true"
    environment:
      NODE_ENV: production
    depends_on:
      - backend
    restart: unless-stopped

  caddy:
    image: caddy:2
    ports:
      - "80:80"
      - "443:443"
    environment:
      # Local default ':80' -> http://localhost. Set to a domain in production
      # for automatic HTTPS (or run plain HTTP behind a Cloudflare Tunnel).
      PROVENANCE_SITE_ADDRESS: ${PROVENANCE_SITE_ADDRESS:-:80}
    volumes:
      - ./Caddyfile:/etc/caddy/Caddyfile:ro
      - caddydata:/data
      - caddyconfig:/config
    depends_on:
      - backend
      - frontend
    restart: unless-stopped

  # Cloudflare Tunnel connector. The tunnel/ingress is configured in the
  # Cloudflare dashboard; this container just connects. One public hostname
  # (e.g. provenance.paul.farm) -> http://caddy:80 is enough, because Caddy
  # does the internal path routing (/ -> frontend, /api + /health -> backend).
  #
  # Opt-in via the "tunnel" profile so local dev doesn't start it. On the lab
  # host set COMPOSE_PROFILES=tunnel so `docker compose up -d` includes it.
  cloudflared:
    image: cloudflare/cloudflared:latest
    restart: unless-stopped
    command: tunnel --no-autoupdate run
    environment:
      TUNNEL_TOKEN: ${CLOUDFLARE_TUNNEL_TOKEN:-}
    depends_on:
      - caddy
    profiles:
      - tunnel

# Auto-deploy is handled by the host's global Watchtower (a single
# nickfedor/watchtower instance watches every container labelled
# `com.centurylinklabs.watchtower.enable=true` across all stacks). The backend
# and frontend carry that label above, so a new :test-main image is pulled and
# the container recreated automatically — no per-stack Watchtower needed.

volumes:
  pgdata:
  miniodata:
  caddydata:
  caddyconfig: