添加API网关后Okteto服务反复重启失败求助

使用Okteto部署项目服务，此前运行正常，添加api-gateway后，其中一个服务出现反复重启后失败的情况。

main-flow.yaml #

name: Build and deliver containers

on: push

jobs:
  customerservice:
    runs-on: ubuntu-latest
    name: Java 17 build, test, deliver customer-service container
    steps:
      - name: Check out repository
        uses: actions/checkout@v3

      - name: Set up JDK 17
        uses: actions/setup-java@v3
        with:
          java-version: 17
          cache: maven
          distribution: adopt

      - name: Build, test and package project
        working-directory: customer-service
        run: mvn package -DskipTests

      - name: Docker login
        run: docker login -u ${{ secrets.DOCKER_USER }} -p ${{ secrets.DOCKER_PASSWORD }}

      - name: Docker Build
        run: docker build -t ${{ secrets.DOCKER_USER }}/customer-service:latest ./customer-service

      - name: Upload container to Docker Hub with Push
        run: docker push ${{ secrets.DOCKER_USER }}/customer-service:latest
  taxiservice:
    runs-on: ubuntu-latest
    name: Java 17 build, test, deliver taxi-service container
    steps:
      - name: Check out repository
        uses: actions/checkout@v3

      - name: Set up JDK 17
        uses: actions/setup-java@v3
        with:
          java-version: 17
          cache: maven
          distribution: adopt

      - name: Build, test and package project
        working-directory: taxi-service
        run: mvn package -DskipTests

      - name: Docker login
        run: docker login -u ${{ secrets.DOCKER_USER }} -p ${{ secrets.DOCKER_PASSWORD }}

      - name: Docker Build
        run: docker build -t ${{ secrets.DOCKER_USER }}/taxi-service:latest ./taxi-service

      - name: Upload container to Docker Hub with Push
        run: docker push ${{ secrets.DOCKER_USER }}/taxi-service:latest
  tripservice:
    runs-on: ubuntu-latest
    name: Java 17 build, test, deliver trip-service container
    steps:
      - name: Check out repository
        uses: actions/checkout@v3

      - name: Set up JDK 17
        uses: actions/setup-java@v3
        with:
          java-version: 17
          cache: maven
          distribution: adopt

      - name: Build, test and package project
        working-directory: trip-service
        run: mvn package -DskipTests

      - name: Docker login
        run: docker login -u ${{ secrets.DOCKER_USER }} -p ${{ secrets.DOCKER_PASSWORD }}

      - name: Docker Build
        run: docker build -t ${{ secrets.DOCKER_USER }}/trip-service:latest ./trip-service

      - name: Upload container to Docker Hub with Push
        run: docker push ${{ secrets.DOCKER_USER }}/trip-service:latest
  apigateway:
    runs-on: ubuntu-latest
    name: Java 17 build, test, deliver ede-api-gateway container
    steps:
      - name: Check out repository
        uses: actions/checkout@v3

      - name: Set up JDK 17
        uses: actions/setup-java@v3
        with:
          java-version: 17
          cache: maven
          distribution: adopt

      - name: Build, test and package project
        working-directory: ./api-gateway
        run: mvn package -DskipTests

      - name: Docker login
        run: docker login -u ${{ secrets.DOCKER_USER }} -p ${{ secrets.DOCKER_PASSWORD }}

      - name: Docker Build
        run: docker build -t ${{ secrets.DOCKER_USER }}/ede-project-api-gateway:latest ./api-gateway

      - name: Upload container to Docker Hub with Push
        run: docker push ${{ secrets.DOCKER_USER }}/ede-project-api-gateway:latest

docker-compose.yaml #

services:
  mongo-customers:
    image: mongo:latest
    ports:
      - 27017:27017
    volumes:
      - mongodb_customer_data:/data/db

  mysql-trip:
    image: mysql:latest
    ports:
      - 3306:3306
    volumes:
      - mysql_trip_data:/var/lib/mysql
    environment:
      MYSQL_ROOT_PASSWORD: trippassword

  mysql-taxi:
    image: mysql:latest
    ports:
      - 3307:3306
    volumes:
      - mysql_taxi_data:/var/lib/mysql
    environment:
      MYSQL_ROOT_PASSWORD: taxipassword

  customer-service:
    depends_on:
      - mongo-customers
    image: compacte/customer-service
    ports:
      - 8081:8081
    links:
      - mongo-customers
    environment:
      MONGODB_PORT: 27017
      MONGODB_HOST: mongo-customers

  taxi-service:
    depends_on:
      - mysql-taxi
    image: compacte/taxi-service
    ports:
      - 8082:8082
    links:
      - mysql-taxi
    environment:
      MYSQL_DB_USERNAME: root
      MYSQL_DB_PASSWORD: taxipassword
      MYSQL_DB_HOST: mysql-taxi
      MYSQL_DB_PORT: 3307

  trip-service:
    depends_on:
      - taxi-service
      - customer-service
    image: compacte/trip-service
    ports:
      - 8083:8083
    links:
      - taxi-service
      - customer-service
    environment:
      CUSTOMER_SERVICE_BASEURL: customer-service:8081
      TAXI_INFO_SERVICE_BASEURL: taxi-service:8082
      MYSQL_DB_USERNAME: root
      MYSQL_DB_PASSWORD: trippassword
      MYSQL_DB_HOST: mysql-trip
      MYSQL_DB_PORT: 3306
    
  api-gateway:
    depends_on:
      - taxi-service
      - trip-service
    image: compacte/ede-project-api-gateway
    ports:
      - 8084:8084
    links:
      - taxi-service
      - trip-service
    environment:
      TAXI_SERVICE_BASEURL: taxi-service:8082
      TRIP_SERVICE_BASEURL: trip-service:8083

volumes:
  mongodb_customer_data:
  mysql_taxi_data:
  mysql_trip_data:

排查建议 #

• 优先查看服务日志：执行okteto logs <失败服务名>或在Okteto控制台查看具体报错信息，比如数据库连接超时、服务调用失败、内存不足等，这是最快定位问题的方式。
• 验证依赖服务就绪状态：depends_on仅保证容器启动顺序，不确保服务可用。例如trip-service依赖的taxi-service可能在MySQL未完成初始化时就启动，导致连接失败。可以给数据库和服务添加健康检查：

  mysql-taxi:
    # 原有配置
    healthcheck:
      test: ["CMD", "mysqladmin", "ping", "-h", "localhost"]
      interval: 10s
      timeout: 5s
      retries: 5
  taxi-service:
    depends_on:
      mysql-taxi:
        condition: service_healthy
  

• 检查资源配额：Okteto存在资源限制，新增api-gateway后可能导致总内存/CPU超出配额，触发OOM杀死服务。前往Okteto控制台查看资源使用情况，可给各服务添加资源限制：

  taxi-service:
    # 原有配置
    resources:
      requests:
        memory: "256Mi"
        cpu: "100m"
      limits:
        memory: "512Mi"
        cpu: "500m"
  

• 核对API网关配置：检查api-gateway的路由规则是否正确，是否存在循环调用或错误转发导致目标服务负载过高。查看api-gateway日志，确认它能正常连接到taxi-service和trip-service。
• 校验环境变量：确认失败服务的环境变量配置，比如trip-service的MySQL端口为容器内部的3306（而非宿主机端口），当前配置看起来正确，但可再次核对。

内容的提问来源于stack exchange，提问作者Seppe