cfc/cfc-langgraph/docker-compose.yml

version: "3.8"

services:
  cfc-backend:
    build:
      context: ../cfc-backend
      dockerfile: Dockerfile
    ports:
      - "9082:9082"
    environment:
      - SPRING_PROFILES_ACTIVE=prod
      - python.enabled=true
      - python.base-url=http://langgraph-svc:9000
    networks:
      - cfc-net
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:9082/actuator/health"]
      interval: 30s
      timeout: 10s
      retries: 3

  langgraph-svc:
    image: cfc-langgraph-langgraph-svc:v10.3
    build:
      context: .
      dockerfile: Dockerfile
    ports:
      - "9000:9000"
    env_file:
      - .env.production
    environment:
      - JAVA_BASE_URL=http://172.17.0.1:9082
      - JAVA_CONTEXT_URL=http://172.17.0.1:9082/api/ai/context
      - CHROMA_DB_PATH=/data/chroma_db
      - JSON_LOGS=true
      # LangSmith 三个变量从 env_file (.env.production) 读取，不再覆盖
    volumes:
      # 共享 Java 文件上传目录（LangGraph 读取已上传的 PDF 文件）
      - /data/cfc-uploads:/data/cfc-uploads
      # Chroma 向量数据库持久化
      - langgraph-data:/data/chroma_db
    networks:
      - cfc-net
    restart: unless-stopped
    # 内存硬限制：防止 ChromaDB/搜索任务异常膨胀吃满宿主机内存导致整机 OOM
    mem_limit: 8g
    healthcheck:
      test: ["CMD", "python", "-c", "import urllib.request,sys; sys.exit(0 if urllib.request.urlopen('http://localhost:9000/health', timeout=5).status==200 else 1)"]
      interval: 30s
      timeout: 10s
      retries: 3


  prometheus:
    image: prom/prometheus:latest
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml
      - prometheus-data:/prometheus
    networks:
      - cfc-net
    restart: unless-stopped
    profiles:
      - monitoring

networks:
  cfc-net:

volumes:
  langgraph-data:
  prometheus-data: