learning_ai_common_plat/docker-compose.yml

services:
  # ── Mailpit SMTP Sandbox (prototype only) ────────────────────
  mailpit:
    image: axllent/mailpit:v1.27.5
    ports:
      - '1025:1025'
      - '8025:8025'
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:8025']
      interval: 10s
      timeout: 5s
      retries: 6
    restart: unless-stopped

  # ── Azurite Blob Storage (prototype only) ─────────────────────
  azurite:
    image: mcr.microsoft.com/azure-storage/azurite:3.35.0
    command: azurite-blob --blobHost 0.0.0.0 --blobPort 10000 --location /data --skipApiVersionCheck
    ports:
      - '10000:10000'
    volumes:
      - azurite-data:/data
    healthcheck:
      test:
        [
          'CMD',
          'node',
          '-e',
          'const net=require("net");const s=net.connect(10000,"127.0.0.1",()=>{s.end();process.exit(0)});s.on("error",()=>process.exit(1));',
        ]
      interval: 10s
      timeout: 5s
      retries: 6
    restart: unless-stopped

  # ── Azure Cosmos DB Emulator (prototype only) ─────────────────
  cosmos-emulator:
    image: mcr.microsoft.com/cosmosdb/linux/azure-cosmos-emulator:vnext-preview
    ports:
      - '8081:8081'
      - '1234:1234'
    environment:
      - PROTOCOL=http
      - ENABLE_EXPLORER=true
      - GATEWAY_PUBLIC_ENDPOINT=cosmos-emulator
    healthcheck:
      test:
        [
          'CMD-SHELL',
          'bash -lc ''exec 3<>/dev/tcp/127.0.0.1/8080; printf "GET /ready HTTP/1.1\r\nHost: localhost\r\nConnection: close\r\n\r\n" >&3; grep -q "200 OK" <&3''',
        ]
      interval: 10s
      timeout: 5s
      retries: 12
      start_period: 20s
    restart: unless-stopped

  # ── Loki (Log Aggregation) ────────────────────────────────────
  loki:
    image: grafana/loki:3.3.2
    ports:
      - '3100:3100'
    volumes:
      - ./services/monitoring/loki/loki-config.yml:/etc/loki/local-config.yaml
      - loki-data:/loki
    command: -config.file=/etc/loki/local-config.yaml
    restart: unless-stopped
    healthcheck:
      # BusyBox wget (used in Alpine images) doesn't support --no-verbose/--tries
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:3100/ready']
      interval: 15s
      timeout: 5s
      retries: 3

  # ── Grafana (Log Viewer + Dashboards) ─────────────────────────
  grafana:
    image: grafana/grafana:11.4.0
    ports:
      - '3000:3000'
    environment:
      - GF_SECURITY_ADMIN_USER=admin
      - GF_SECURITY_ADMIN_PASSWORD=lysnrai
      - GF_USERS_ALLOW_SIGN_UP=false
    volumes:
      - ./services/monitoring/grafana/provisioning:/etc/grafana/provisioning
      - ./services/monitoring/grafana/dashboards:/var/lib/grafana/dashboards
      - grafana-data:/var/lib/grafana
    depends_on:
      loki:
        condition: service_started
    restart: unless-stopped
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:3000/api/health']
      interval: 15s
      timeout: 5s
      retries: 3

  # ── API Gateway (Traefik) ───────────────────────────────────
  gateway:
    image: traefik:v3.3
    command:
      - '--api.insecure=true'
      - '--providers.docker=true'
      - '--providers.docker.exposedbydefault=false'
      - '--entrypoints.web.address=:80'
      - '--accesslog=true'
      - '--accesslog.format=json'
    ports:
      - '80:80'
      - '8080:8080' # Traefik dashboard
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock:ro
    depends_on:
      loki:
        condition: service_started
    restart: unless-stopped

  # ── Platform Service (Fastify + TypeScript) ─────────────
  # Consolidated: auth, audit, notifications, flags, blob, invitations, referrals, promos,
  #               subscriptions, usage, plans, licenses, stripe, items, comments, votes, public
  platform-service:
    build:
      context: .
      dockerfile: services/platform-service/Dockerfile
    ports:
      - '4003:4003'
    env_file:
      - .env
    environment:
      - PORT=4003
      # Local/dev convenience: ensure Cosmos DB + containers exist.
      - COSMOS_AUTO_INIT=true
      - PLATFORM_SERVICE_URL=http://platform-service:4003
      - EXTRACTION_SERVICE_URL=http://extraction-service:4005
      - MCP_SERVER_URL=http://mcp-server:4007
      - MAILPIT_UI_URL=http://mailpit:8025
    depends_on:
      mailpit:
        condition: service_healthy
      azurite:
        condition: service_healthy
      cosmos-emulator:
        condition: service_healthy
    labels:
      - 'traefik.enable=true'
      - 'traefik.http.routers.platform.rule=PathPrefix(`/api`) || PathPrefix(`/public`) || PathPrefix(`/health`)'
      - 'traefik.http.services.platform.loadbalancer.server.port=4003'
    restart: unless-stopped
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:4003/health']
      interval: 30s
      timeout: 10s
      retries: 3

  # ── Extraction Service (Fastify + TypeScript + Python sidecar) ──
  extraction-service:
    build:
      context: .
      dockerfile: services/extraction-service/Dockerfile
    ports:
      - '4005:4005'
    env_file:
      - .env
    environment:
      - PORT=4005
      - PYTHON_SIDECAR_URL=http://localhost:4006
    depends_on:
      cosmos-emulator:
        condition: service_healthy
    labels:
      - 'traefik.enable=true'
      - 'traefik.http.routers.extraction.rule=PathPrefix(`/api/extract`) || PathPrefix(`/api/tasks`)'
      - 'traefik.http.services.extraction.loadbalancer.server.port=4005'
    restart: unless-stopped
    healthcheck:
      test:
        [
          'CMD',
          'node',
          '-e',
          'fetch("http://127.0.0.1:4005/health").then(r => process.exit(r.ok ? 0 : 1)).catch(() => process.exit(1))',
        ]
      interval: 30s
      timeout: 10s
      retries: 3

  # ── MCP Server (Fastify + TypeScript) ────────────────────────
  # Exposes tool namespaces: platform.telemetry.*, platform.diagnostics.*,
  # extraction.*, support.* — consumed by AI agents + admin tooling
  mcp-server:
    build:
      context: .
      dockerfile: services/mcp-server/Dockerfile
    ports:
      - '4007:4007'
    env_file:
      - .env
    environment:
      - PORT=4007
      - PLATFORM_SERVICE_URL=http://platform-service:4003
      - EXTRACTION_SERVICE_URL=http://extraction-service:4005
    depends_on:
      platform-service:
        condition: service_healthy
      extraction-service:
        condition: service_healthy
    labels:
      - 'traefik.enable=true'
      - 'traefik.http.routers.mcp.rule=PathPrefix(`/api/tools`)'
      - 'traefik.http.services.mcp.loadbalancer.server.port=4007'
    restart: unless-stopped
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:4007/health']
      interval: 30s
      timeout: 10s
      retries: 3

  # ── Cowork Service (Fastify bridge to Rust agent runtime) ──────
  # Bridges Tauri desktop / clients and cowork-orchestrator, delegating auth,
  # flags, audit, telemetry, and AI budgets to platform-service.
  cowork-service:
    build:
      context: .
      dockerfile: services/cowork-service/Dockerfile
    ports:
      - '4009:4009'
    env_file:
      - .env
    environment:
      - PORT=4009
      - NODE_ENV=development
      - PRODUCT_ID=clawcowork
      - COSMOS_ENDPOINT=https://cosmos-emulator:8081
      - PLATFORM_SERVICE_URL=http://platform-service:4003
      - EXTRACTION_SERVICE_URL=http://extraction-service:4005
    depends_on:
      cosmos-emulator:
        condition: service_healthy
      platform-service:
        condition: service_healthy
    labels:
      - 'traefik.enable=true'
      - 'traefik.http.routers.cowork.rule=PathPrefix(`/api/cowork`)'
      - 'traefik.http.services.cowork.loadbalancer.server.port=4009'
    restart: unless-stopped
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:4009/health']
      interval: 10s
      timeout: 5s
      retries: 10

  # ── Admin Web (Next.js Platform Admin Console) ─────────────────
  admin-web:
    build:
      context: .
      dockerfile: dashboards/admin-web/Dockerfile
      args:
        BYTELYST_PACKAGE_SOURCE: ${BYTELYST_PACKAGE_SOURCE:-vendor}
        NEXT_PUBLIC_PRODUCT_ID: ${NEXT_PUBLIC_PRODUCT_ID:-admin}
        NEXT_PUBLIC_PLATFORM_URL: http://platform-service:4003
    container_name: admin-web
    ports:
      - '3001:3001'
    networks:
      - default
    restart: unless-stopped
    depends_on:
      platform-service:
        condition: service_healthy
    environment:
      - NODE_ENV=production
    healthcheck:
      test: ['CMD', 'wget', '-q', '--spider', 'http://127.0.0.1:3001']
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 15s

# ── Volumes ───────────────────────────────────────────────────────
volumes:
  azurite-data:
  loki-data:
  grafana-data: