【Docker系列】Docker Compose 服务依赖和健康检查

当依赖的服务，启动失败或者错误时，后续的服务哪怕成功启动，业务无法运行的。这时就需要健康检查！健康检查是容器运行状态的高级检查，主要是检查容器所运行的进程是否能正常的对外提供“服务”，比如一个数据库容器，我们不光需要这个容器是up的状态，我们还要求这个容器的数据库进程能够正常对外提供服务，这就是所谓的健康检查。

小叶柏杉

8249人浏览 · 2022-01-22 05:54:57

小叶柏杉 · 2022-01-22 05:54:57 发布

准备

不想再写一遍了，请看上篇文章的文件准备：【Docker系列】Docker Compose 环境变量

服务依赖

docker-compose.yml 添加depends_on参数

启动顺序：

redis-server
flask
nginx

version: "3.8"

services:
  flask:
    build:
      context: ./flask
      dockerfile: Dockerfile
    image: flask-demo:latest
    environment:
      - REDIS_HOST=redis-server
      - REDIS_PASS=${REDIS_PASSWORD}
    depends_on: # 依赖redis-server启动后，再启动
      - redis-server 
    networks:
      - backend
      - frontend

  redis-server:
    image: redis:latest
    command: redis-server --requirepass ${REDIS_PASSWORD}
    networks:
      - backend

  nginx:
    image: nginx:stable-alpine
    ports:
      - 8000:80
    depends_on: # 依赖flask启动后，再启动
      - flask
    volumes:
      - ./nginx/nginx.conf:/etc/nginx/conf.d/default.conf:ro
      - ./var/log/nginx:/var/log/nginx
    networks:
      - frontend

networks:
  backend:
  frontend:

健康检查

当依赖的服务，启动失败或者错误时，后续的服务哪怕成功启动，业务无法运行的。这时就需要健康检查！

健康检查是容器运行状态的高级检查，主要是检查容器所运行的进程是否能正常的对外提供“服务”，比如一个数据库容器，我们不光需要这个容器是up的状态，我们还要求这个容器的数据库进程能够正常对外提供服务，这就是所谓的健康检查。

容器本身有一个健康检查的功能，但是需要在Dockerfile里定义，或者在执行docker container run 的时候，通过下面的一些参数指定

--health-cmd string              Command to run to check health
--health-interval duration       Time between running the check
                                (ms|s|m|h) (default 0s)
--health-retries int             Consecutive failures needed to
                                report unhealthy
--health-start-period duration   Start period for the container to
                                initialize before starting
                                health-retries countdown
                                (ms|s|m|h) (default 0s)
--health-timeout duration        Maximum time to allow one check to

HEALTHCHECK

官网参数说明：https://docs.docker.com/engine/reference/builder/#healthcheck

可以在Dockerfile里面，写健康检测的命令

Dockerfile文件，添加一行HEALTHCHECK

python:3.9.5-slim镜像里面，记得安装一下 curl命令

FROM python:3.9.5-slim

RUN pip install flask redis && \
    apt-get update && \
    apt-get install -y curl && \  # 需要安装一下curl
    groupadd -r flask && useradd -r -g flask flask && \
    mkdir /src && \
    chown -R flask:flask /src

USER flask

COPY app.py /src/app.py

WORKDIR /src

ENV FLASK_APP=app.py REDIS_HOST=redis

EXPOSE 5000

HEALTHCHECK --interval=30s --timeout=3s \
    CMD curl -f http://localhost:5000/ || exit 1

CMD ["flask", "run", "-h", "0.0.0.0"]

会每隔30秒检查一次，如果失败就会退出，退出代码是1

HEALTHCHECK --interval=30s --timeout=3s \
    CMD curl -f http://localhost:5000/ || exit 1

构建镜像和创建容器

构建镜像，创建一个bridge网络，然后启动容器连到bridge网络

设计场景：仅启动flask-demo容器，依赖的redis服务不启动，flask将无法成功连接redis数据库。

$ docker image build -t flask-demo .
$ docker network create mybridge
$ docker container run -d --network mybridge --env REDIS_PASS=abc123 flask-demo

查看容器状态

$ docker container ls
CONTAINER ID   IMAGE        COMMAND                  CREATED       STATUS                            PORTS      NAMES
059c12486019   flask-demo   "flask run -h 0.0.0.0"   4 hours ago   Up 8 seconds (health: starting)   5000/tcp   dazzling_tereshkov

When a container has a healthcheck specified, it has a health status in addition to its normal status. This status is initially starting. Whenever a health check passes, it becomes healthy (whatever state it was previously in). After a certain number of consecutive failures, it becomes unhealthy.
当容器指定了运行状况检查时，它除了正常状态外还具有运行状况。此状态处于starting.。只要健康检查通过，它就会变得healthy（无论以前处于何种状态）。经过一定数量的连续失败后，它变得unhealthy。

The options that can appear before CMD are:

--interval=DURATION (default: 30s)
--timeout=DURATION (default: 30s)
--start-period=DURATION (default: 0s)
--retries=N (default: 3) 默认3次

也可以通过docker container inspect 059 查看详情，其中有有关health的

"Health": {
"Status": "starting",
"FailingStreak": 1,
"Log": [
    {
        "Start": "2021-07-14T19:04:46.4054004Z",
        "End": "2021-07-14T19:04:49.4055393Z",
        "ExitCode": -1, # 健康检查码，成功返回 0
        "Output": "Health check exceeded timeout (3s)"
    }
]
}

经过3次检查，一直是不通的，然后health的状态会从starting变为 unhealthy

docker container ls
CONTAINER ID   IMAGE        COMMAND                  CREATED       STATUS                     PORTS      NAMES
059c12486019   flask-demo   "flask run -h 0.0.0.0"   4 hours ago   Up 2 minutes (unhealthy)   5000/tcp   dazzling_tereshkova

启动redis服务器

启动redis，连到mybridge上，--name=redis，注意密码 --requirepass abc123

$ docker container run -d --network mybridge --name redis redis:latest redis-server --requirepass abc123

经过几秒钟，我们的 flask 变成了healthy

$ docker container ls
CONTAINER ID   IMAGE          COMMAND                  CREATED          STATUS                   PORTS      NAMES
bc4e826ee938   redis:latest   "docker-entrypoint.s…"   18 seconds ago   Up 16 seconds            6379/tcp   redis
059c12486019   flask-demo     "flask run -h 0.0.0.0"   4 hours ago      Up 6 minutes (healthy)   5000/tcp   dazzling_tereshkova

docker-compose 健康检查 (重点)

官网示例：

healthcheck:
  test: ["CMD", "curl", "-f", "http://localhost"]
  interval: 1m30s
  timeout: 10s
  retries: 3
  start_period: 40s

interval, timeout 和start_period 指定为持续时间

docker-compose.yml

version: "3.8"

services:
  flask:
    build:
      context: ./flask
      dockerfile: Dockerfile
    image: flask-demo:latest
    environment:
      - REDIS_HOST=redis-server
      - REDIS_PASS=${REDIS_PASSWORD}
    healthcheck: # 添加健康检测
      test: ["CMD", "curl", "-f", "http://localhost:5000"]
      interval: 30s
      timeout: 3s
      retries: 3
      start_period: 40s
    depends_on:
      redis-server:
        condition: service_healthy
    networks:
      - backend
      - frontend

  redis-server:
    image: redis:latest
    command: redis-server --requirepass ${REDIS_PASSWORD}
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 1s
      timeout: 3s
      retries: 30
    networks:
      - backend

  nginx:
    image: nginx:stable-alpine
    ports:
      - 8000:80
    depends_on:
      - flask
    volumes:
      - ./nginx/nginx.conf:/etc/nginx/conf.d/default.conf:ro
      - ./var/log/nginx:/var/log/nginx
    networks:
      - frontend

networks:
  backend:
  frontend:

为什么仅有flask容器健康检测了？

PS C:\Users\柏杉\Downloads\compose-healthcheck-flask> docker-compose ps
                  Name                                Command                       State                  Ports
------------------------------------------------------------------------------------------------------------------------
compose-healthcheck-flask_flask_1          flask run -h 0.0.0.0             Up (health: starting)   5000/tcp
compose-healthcheck-flask_nginx_1          /docker-entrypoint.sh ngin ...   Up                      0.0.0.0:8000->80/tcp
compose-healthcheck-flask_redis-server_1   docker-entrypoint.sh redis ...   Up                      6379/tcp

在docker-compose.yml文件里面，去定义的Nginx是depends_on，没有healthcheck健康检测，

nginx 依赖的flask的状态是healthy的时候，才会启动，否则等待。

  nginx:
    image: nginx:stable-alpine
    ports:
      - 8000:80
    depends_on:
      flask:
        condition: service_healthy
    volumes:
      - ./nginx/nginx.conf:/etc/nginx/conf.d/default.conf:ro
      - ./var/log/nginx:/var/log/nginx
    networks:
      - frontend

执行docker-compose up -d会卡主，因为flask不是healthy状态！

redis-server 添加一个healthcheck

  redis-server:
    image: redis:latest
    command: redis-server --requirepass ${REDIS_PASSWORD}
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 1s
      timeout: 3s
      retries: 30
    networks:
      - backend