autollama/docker-compose.yaml

Modern JavaScript-first RAG framework with contextual embeddings, professional CLI, and one-command deployment

services:
  postgres:
    image: postgres:15-alpine
    container_name: autollama-postgres
    environment:
      POSTGRES_USER: autollama
      POSTGRES_PASSWORD: autollama
      POSTGRES_DB: autollama
      # Memory optimizations for small instances
      POSTGRES_SHARED_BUFFERS: 64MB
      POSTGRES_EFFECTIVE_CACHE_SIZE: 512MB
      POSTGRES_WORK_MEM: 2MB
      POSTGRES_MAINTENANCE_WORK_MEM: 32MB
    volumes:
      - postgres_data:/var/lib/postgresql/data
    ports:
      - "5432:5432"
    networks:
      - autollama-network
    deploy:
      resources:
        limits:
          memory: 256M
        reservations:
          memory: 128M
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U autollama"]
      interval: 10s
      retries: 5
    restart: unless-stopped

  qdrant:
    image: qdrant/qdrant:latest
    container_name: autollama-qdrant
    environment:
      # Memory optimizations for Qdrant
      QDRANT__SERVICE__MAX_REQUEST_SIZE_MB: 32
      QDRANT__STORAGE__PERFORMANCE__MAX_SEARCH_THREADS: 2
      QDRANT__STORAGE__OPTIMIZERS__MEMMAP_THRESHOLD: 100000
    ports:
      - "6333:6333"
    volumes:
      - qdrant_data:/qdrant/storage
    networks:
      - autollama-network
    deploy:
      resources:
        limits:
          memory: 512M
        reservations:
          memory: 256M
    restart: unless-stopped

  autollama-bm25:
    build: ./bm25-service
    container_name: autollama-bm25
    environment:
      PORT: 3002
      PYTHONPATH: /app
    volumes:
      - autollama-bm25-data:/app/data/bm25_indices
    networks:
      - autollama-network
    ports:
      - "3012:3002"
    deploy:
      resources:
        limits:
          memory: 256M
        reservations:
          memory: 128M
    restart: unless-stopped

  autollama-api:
    build: ./api
    container_name: autollama-api
    env_file:
      - .env
    environment:
      NODE_ENV: production
      PORT: 3001
      DATABASE_URL: postgresql://autollama:autollama@postgres:5432/autollama
      QDRANT_URL: http://autollama-qdrant:6333
      # Auto-migration settings
      AUTO_MIGRATE: "true"
      MIGRATION_ON_START: "true"
      DB_HOST: postgres
      DB_PORT: 5432
      DB_USER: autollama
      DB_NAME: autollama
      # Node.js memory optimizations
      NODE_OPTIONS: "--max-old-space-size=512 --max-semi-space-size=64"
    networks:
      - autollama-network
    ports:
      - "3001:3001"
      - "3003:3003"
    depends_on:
      postgres:
        condition: service_healthy
      qdrant:
        condition: service_started
      autollama-bm25:
        condition: service_started
    deploy:
      resources:
        limits:
          memory: 1G
        reservations:
          memory: 512M
    restart: unless-stopped

  autollama-frontend:
    build:
      context: ./config/react-frontend
      dockerfile: Dockerfile
    container_name: autollama-frontend
    environment:
      NODE_ENV: production
    networks:
      - autollama-network
    ports:
      - "8080:80"
    depends_on:
      - autollama-api
    deploy:
      resources:
        limits:
          memory: 128M
        reservations:
          memory: 64M
    restart: unless-stopped

volumes:
  postgres_data:
  qdrant_data:
  autollama-bm25-data:

networks:
  autollama-network:
    driver: bridge
    # All services communicate via service names (e.g., autollama-api:3001)
    # DO NOT use localhost in nginx proxy configs - use service names instead