11.cloud与docker常见问题总结

小灰灰 2023-07-04 14:19 3阅读 0赞

# 1. eureka注册慢 #

服务注册到eureka的过程较慢，可能需要发送3次心跳

默认30秒一次心跳，client像server发送心跳的间隔

eureka:
      client:
        serviceUrl:
          defaultZone: http://localhost:8761/eureka/
      instance:
        lease-renewal-interval-in-seconds: 30

# 2. 已经停止的微服务节点注销慢 或 不注销 #

eureka server清理无效节点周期长 （默认90秒）

eureka:
      server:
        enable-self-preservation: false # 关闭eureka的自我保护
        eviction-interval-timer-in-ms: 60000 #清理间隔，默认是60秒，单位毫秒，可以改成4000

eureka client的配置

eureka:
      client:
        healthcheck:
          enabled: true #开启健康检查
        registerWithEureka: false
        fetchRegistry: false
        serviceUrl:
          defaultZone: http://localhost:8761/eureka/
          #http://${eureka.instance.hostname}:${server.port}/eureka/
      instance:
        hostname: localhost
        lease-renewal-interval-in-seconds: 30 #续约更新时间间隔 默认30秒,改成10秒
        lease-expiration-duration-in-seconds: 90 #续约到期时间 默认90秒，改成30秒

# 自定义微服务的 Instance ID #

instance Id 唯一标识注册到 eureka server上的 微服务实例

默认为： s p r i n g . c l o u d . h o s t n a m e : \{spring.cloud.hostname\}: spring.cloud.hostname:\{spring.application.name\}:

s p r i n g . a p p l i c a t i o n . i n s t a n c e i d : \{spring.application.instance\_id\}: spring.application.instanceid:\{server.port\}

eureka:
      instance:
        instance-id: ${ spring.cloud.client.ipAddress}:${ server.port}

# eureka unknown #

1.  应用程序的unknown
    
    1.      spring:
              application:
                name: microservice-provider-user #未配置
    2.  
2.  应用状态 的 unknown
    
        eureka:
          client:
            healthcheck:
              enabled: true # 配置到 application.yml中

# hystrix feign 首次请求失败 #

hystrix 默认超时时间是 1秒，超时就会进入 fallback逻辑

由于spring懒加载，首次比较慢

hystrix:
      command:
        default:
          execution:
            timeout:
              enabled: false # 禁用hystrix
            isolation:
              thread:
                timeoutInMilliseconds: 5000 #延长feign超时为5秒
    feign:
      hystrix:
        enabled: false #禁用hystrix，比较极端

# turbine 聚合的数据不完整 #

turbine:
      appConfig: microservice-consumer-movie,microservice-consumer-movie-feign-hystrix-fallback-stream
      clusterNameExpression: "'default'"

turbine聚合在同一台主机上，就会出现改问题

eureka:
      instance:
        virtual-host-name: ribbon #配置hostname

turbine:
      appConfig: microservice-consumer-movie,microservice-consumer-movie-feign-hystrix-fallback-stream
      clusterNameExpression: "'default'"
      combine-host-port: true #设置为true,新版默认是ture