Hystrix参数踩坑

先来看两个本次我们项目当中出现的异常（出异常时系统未配置hystrix参数）

Caused by: java.util.concurrent.RejectedExecutionException: Task java.util.concurrent.FutureTask@28ee4bd7 rejected from java.util.concurrent.ThreadPoolExecutor@58ab318d[Running, pool size = 10, active threads = 10, queued tasks = 0, completed tasks = 143639]
        at java.util.concurrent.ThreadPoolExecutor$AbortPolicy.rejectedExecution(ThreadPoolExecutor.java:2063)
        at java.util.concurrent.ThreadPoolExecutor.reject(ThreadPoolExecutor.java:830)
        at java.util.concurrent.ThreadPoolExecutor.execute(ThreadPoolExecutor.java:1379)
        at java.util.concurrent.AbstractExecutorService.submit(AbstractExecutorService.java:112)
        at com.netflix.hystrix.strategy.concurrency.HystrixContextScheduler$ThreadPoolWorker.schedule(HystrixContextScheduler.java:172)
        at com.netflix.hystrix.strategy.concurrency.HystrixContextScheduler$HystrixContextSchedulerWorker.schedule(HystrixContextScheduler.java:106)
        at rx.internal.operators.OperatorSubscribeOn.call(OperatorSubscribeOn.java:50)
        at rx.internal.operators.OperatorSubscribeOn.call(OperatorSubscribeOn.java:30)
        at rx.internal.operators.OnSubscribeLift.call(OnSubscribeLift.java:48)
        ... 161 common frames omitted

Exception in thread "Thread-23" com.netflix.hystrix.exception.HystrixRuntimeException: OrderQueryClient#orderDetail(String) fallback execution rejected.

异常原因

解释下这两个异常出现的原因

RejectedExecutionException

hystrix 默认的策略模式为 THREAD，并且默认核心线程为10，其队列默认使用 SynchronousQueue，也就是最大并发只能到10，超过则直接拒绝，然后走降级逻辑

HystrixRuntimeException：fallback execution rejected.

这个异常表明 降级逻辑执行被拒绝，hystrix.command.default.fallback.isolation.semaphore.maxConcurrentRequests 默认值为10，表明降级执行超过10并发，则会抛出此异常

Hystrix配置链接

修改参数配置

# 参考配置如下
hystrix:
  # === === === == 默认Command === === === ==
  command:
    default:
      # 降级信号量控制
      fallback:
        enable: true
        isolation:
          semaphore:
            maxConcurrentRequests: 200
      execution:
        isolation:
          # 调用隔离方式, 默认: 采用线程隔离, ExecutionIsolationStrategy:THREAD
          strategy: THREAD
          # 调用超时时间, 默认: 5 秒
          thread:
            timeoutInMilliseconds: 30000
      # === === === == 熔断器 === === === ==
      circuitBreaker:
        # 熔断器在整个统计时间内是否开启的阀值, 默认20个请求
        requestVolumeThreshold: 1000
        # 熔断器默认工作时间, 默认: 5 秒
        sleepWindowInMilliseconds: 5
        # 默认: 50%, 当出错率超过50% 后熔断器启动
        errorThresholdPercentage: 50
        # 是否强制开启熔断器阻断所有请求, 默认: false, 不开启
        forceOpen: false
        # 是否允许熔断器忽略错误, 默认false, 不开启
        forceClosed: false
  threadpool:
    default:
      # 并发执行的最大线程数，默认10
      coreSize: 400 
      # SynchronousQueue
      maxQueueSize: -1 
      # 线程生存时间, 默认1分钟
      keepAliveTimeMinutes: 2

后记思考

hystrix 的 Isolation 策略默认是服务间隔离，隔离方式有两种：线程池、信号量

默认是采用线程池隔离，从而某个服务出现问题时，不会影响其他服务（底层执行只会命中对应的池然后池溢出降级）

但是此时我们项目的场景比较单一，它仅仅是两个单体项目之间的调用：A->B，其中涉及非常多的接口，而我们微观来看，实际上也不想让某些“非核心”接口，影响了其他核心接口的调用，这其实就需要hystrix提供更细粒度的断路器配置

Hystrix它是如何工作的