ai_agent/docker-compose.prod.yml

# Production configuration
# without reverse proxy (ports exposed directly)
#
# Usage:
#   1. Fill backend/.env with real production secrets (cp backend/.env.example
#      backend/.env if it doesn't exist yet). The same file is used for dev and
#      prod — there is no separate .env.prod.
#   2. Run: make prod   (or: docker compose --env-file backend/.env -f docker-compose.prod.yml up -d)

services:

  app:
    build:
      context: ./backend
      dockerfile: Dockerfile
    container_name: ai_agent_backend
    volumes:
      - media_data:/app/media
    env_file:
      - ./backend/.env
    environment:
      - DEBUG=false
      - ENVIRONMENT=production
      - POSTGRES_HOST=db
      - REDIS_HOST=redis
      - MILVUS_HOST=milvus
      - MILVUS_PORT=19530
      - TASKIQ_BROKER_URL=redis://redis:6379/1
      - TASKIQ_RESULT_BACKEND=redis://redis:6379/1
    command: uvicorn app.main:app --host 0.0.0.0 --port 8033 --workers 4
    networks:
      - backend-internal
    ports:
      - "8033:8033"
    depends_on:
      db:
        condition: service_healthy
      redis:
        condition: service_healthy
      milvus:
        condition: service_healthy
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '1'
          memory: 512M
        reservations:
          cpus: '0.5'
          memory: 256M

  db:
    image: postgres:16-alpine
    container_name: ai_agent_db
    environment:
      - POSTGRES_USER=${POSTGRES_USER:-postgres}
      - POSTGRES_PASSWORD=${POSTGRES_PASSWORD}
      - POSTGRES_DB=${POSTGRES_DB:-ai_agent}
    volumes:
      - postgres_data:/var/lib/postgresql/data
    # NO external ports - only accessible within backend-internal network
    networks:
      - backend-internal
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-postgres}"]
      interval: 10s
      timeout: 5s
      retries: 5
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '1'
          memory: 512M

  redis:
    image: redis:7-alpine
    container_name: ai_agent_redis
    command: redis-server --requirepass ${REDIS_PASSWORD:?REDIS_PASSWORD is required}
    # NO external ports - only accessible within backend-internal network
    volumes:
      - redis_data:/data
    networks:
      - backend-internal
    healthcheck:
      test: ["CMD", "redis-cli", "-a", "${REDIS_PASSWORD}", "ping"]
      interval: 10s
      timeout: 5s
      retries: 5
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 256M

  etcd:
    image: quay.io/coreos/etcd:v3.6.11
    container_name: ai_agent_etcd
    volumes:
      - etcd_data:/var/lib/etcd
    networks:
      - backend-internal
    healthcheck:
      test: ["CMD", "etcdctl", "endpoint", "health"]
      interval: 10s
      timeout: 5s
      retries: 5
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '0.25'
          memory: 128M

  minio:
    image: minio/minio:RELEASE.2024-09-22T00-33-43Z
    container_name: ai_agent_minio
    command: server /data --console-address ":9001"
    # NO external ports - only accessible within backend-internal network
    volumes:
      - minio_data:/data
    environment:
      - MINIO_ROOT_USER=${MINIO_ROOT_USER:-minioadmin}
      - MINIO_ROOT_PASSWORD=${MINIO_ROOT_PASSWORD:-changeme}
    networks:
      - backend-internal
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
      interval: 30s
      timeout: 10s
      retries: 5
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 256M

  milvus:
    image: milvusdb/milvus:v2.6.17
    container_name: ai_agent_milvus
    command: ["milvus", "run", "standalone"]
    # NO external ports - only accessible within backend-internal network
    volumes:
      - milvus_data:/var/lib/milvus
    environment:
      - ETCD_ENDPOINTS=etcd:2379
      - MINIO_ADDRESS=minio:9000
    depends_on:
      etcd:
        condition: service_healthy
      minio:
        condition: service_healthy
    networks:
      - backend-internal
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:9091/healthz"]
      # Milvus standalone needs querycoord/datacoord/proxy to converge before
      # /healthz returns 200 — first boot can take 60-180s on Apple Silicon.
      start_period: 90s
      interval: 30s
      timeout: 10s
      retries: 10
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '1'
          memory: 1G

  taskiq_worker:
    image: ai_agent_backend:dev
    container_name: ai_agent_taskiq_worker
    volumes:
      - media_data:/app/media
    command: taskiq worker app.worker.taskiq_app:broker --workers 4
    env_file:
      - ./backend/.env
    environment:
      - DEBUG=false
      - POSTGRES_HOST=db
      - REDIS_HOST=redis
      - MILVUS_HOST=milvus
      - MILVUS_PORT=19530
      - TASKIQ_BROKER_URL=redis://:${REDIS_PASSWORD}@redis:6379/1
      - TASKIQ_RESULT_BACKEND=redis://:${REDIS_PASSWORD}@redis:6379/1
    networks:
      - backend-internal
    depends_on:
      app:
        condition: service_started
      redis:
        condition: service_healthy
      db:
        condition: service_healthy
      milvus:
        condition: service_healthy
    restart: unless-stopped
    deploy:
      replicas: 2
      resources:
        limits:
          cpus: '0.5'
          memory: 512M
        reservations:
          cpus: '0.25'
          memory: 256M

  taskiq_scheduler:
    image: ai_agent_backend:dev
    container_name: ai_agent_taskiq_scheduler
    command: taskiq scheduler app.worker.taskiq_app:scheduler
    env_file:
      - ./backend/.env
    environment:
      - DEBUG=false
      - REDIS_HOST=redis
      - MILVUS_HOST=milvus
      - MILVUS_PORT=19530
      - TASKIQ_BROKER_URL=redis://:${REDIS_PASSWORD}@redis:6379/1
      - TASKIQ_RESULT_BACKEND=redis://:${REDIS_PASSWORD}@redis:6379/1
    networks:
      - backend-internal
    depends_on:
      app:
        condition: service_started
      redis:
        condition: service_healthy
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '0.25'
          memory: 128M

  frontend:
    build:
      context: ./frontend
      dockerfile: Dockerfile
    container_name: ai_agent_frontend
    environment:
      - NODE_ENV=production
      - BACKEND_URL=http://app:8033
      - BACKEND_WS_URL=ws://app:8033
    networks:
      - backend-internal
    ports:
      - "3033:3033"
    depends_on:
      - app
    restart: unless-stopped
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 256M

networks:
  backend-internal:
    driver: bridge
    internal: true

volumes:
  media_data:
  postgres_data:
  redis_data:
  etcd_data:
  minio_data:
  milvus_data: