Docker Compose部署下Celery Worker/Beat无法连接RabbitMQ问题

我尝试通过Docker Compose整合服务，但Celery Worker与Celery Beat服务无法连接RabbitMQ服务。以下是相关配置与错误信息：

RabbitMQ服务配置（docker-compose.yml） #

rabbitmq:
    container_name: "rabbitmq"
    image: rabbitmq:3-management
    ports:
      - 15672:15672
      - 5672:5672
    environment:
      - RABBITMQ_DEFAULT_USER=guest
      - RABBITMQ_DEFAULT_PASS=guest
    depends_on:
      - server
    volumes:
      - rabbitmq:/var/lib/rabbitmq

Celery Worker与Celery Beat服务配置（docker-compose.yml） #

celery_worker:
    container_name: celery-worker
    build: .
    command: celery -A tasks worker -E --loglevel=INFO
    environment:
      host_server: postgresqldb
      db_server_port: 5432
      database_name: db
      db_username: user
      db_password: password
      ssl_mode: prefer
    networks:
      - postgresqlnet
    depends_on:
      - rabbitmq

celery_beat:
    container_name: celery-beat
    build: .
    command: celery -A tasks beat
    environment:
      - host_server=postgresqldb
      - db_server_port=5432
      - database_name=db
      - db_username=user
      - db_password=password
      - ssl_mode=prefer
    networks:
      - postgresqlnet
    depends_on:
      - rabbitmq

celeryconfig.py配置 #

broker_url = "amqp://guest:guest@localhost:5672//"

错误输出 #

celery-worker     | [2023-02-03 14:24:43,223: ERROR/MainProcess] consumer: Cannot connect to amqp://guest:**@127.0.0.1:5672//: [Errno 111] Connection refused.
celery-worker     | Trying again in 32.00 seconds... (16/100)
celery-worker     | 
celery-beat       | [2023-02-03 14:25:10,058: ERROR/MainProcess] beat: Connection error: [Errno 111] Connection refused. Trying again in 32.0 seconds...

解决方法 #

1. 修正RabbitMQ连接地址 #

Docker容器内的localhost指向容器自身，而非RabbitMQ容器。Docker Compose中可直接通过服务名作为域名访问其他容器，因此修改celeryconfig.py的broker_url：

broker_url = "amqp://guest:guest@rabbitmq:5672//"

2. 确保所有服务在同一网络 #

当前Celery服务加入了postgresqlnet网络，但RabbitMQ未配置该网络，导致无法互通。给RabbitMQ服务添加网络配置：

rabbitmq:
    # ... 保留原有配置
    networks:
      - postgresqlnet

3. 添加RabbitMQ健康检查（可选但推荐） #

depends_on仅保证启动顺序，不确保RabbitMQ完全就绪。添加健康检查，让Celery等待RabbitMQ启动完成后再启动：

rabbitmq:
    # ... 保留原有配置
    healthcheck:
      test: ["CMD", "rabbitmq-diagnostics", "-q", "ping"]
      interval: 30s
      timeout: 30s
      retries: 3

celery_worker:
    # ... 保留原有配置
    depends_on:
      rabbitmq:
        condition: service_healthy

celery_beat:
    # ... 保留原有配置
    depends_on:
      rabbitmq:
        condition: service_healthy

内容的提问来源于stack exchange，提问作者redd