现象：

昨天突然线上很多接口获取失败，通过 kibana发现大量异常，具体异常信息：

...into fallback. Rejected command because thread-pool queueSize is at rejection threshold.

异常代码出处：

@FeignClient(name = "api", fallbackFactory = LoadBalancingFallbackFactory.class)public interface LoadBalancingFeignClient {  @PostMapping(value = "/api/loadBalancing/server")  Result currentServer();}@Slf4j@Componentpublic class LoadBalancingFallbackFactory implements FallbackFactory {  @Override  public LoadBalancingFeignClient create(Throwable throwable) {    final String msg = throwable.getMessage();    return () -> {      log.error("loadBalancingFeignClient currentServer into fallback. {}", msg);      return Result.error();    };****  }}

原因：

看到这里已经很明显了，是由于hystrix线程池不够用，直接熔断导致的。项目apollo配置：

hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds = 3500hystrix.threadpool.default.maxQueueSize = 60hystrix.threadpool.default.queueSizeRejectionThreshold = 40

hystrix参数简析：

maxQueueSize：线程池大小，默认为-1，创建的队列是SynchronousQueue，如果设置大于0则根据其大小创建LinkedBlockingQueue。

queueSizeRejectionThreshold：动态控制线程池队列的上限，即使maxQueueSize没有达到，达到queueSizeRejectionThreshold该值后，请求也会被拒绝，默认值5

相关源码：

hystrix-core-1.5.12-sources.jar!/com/netflix/hystrix/strategy/concurrency/HystrixContextScheduler.java

private class HystrixContextSchedulerWorker extends Worker {    private final Worker worker;    private HystrixContextSchedulerWorker(Worker actualWorker) {      this.worker = actualWorker;    }    @Override    public void unsubscribe() {      worker.unsubscribe();    }    @Override    public boolean isUnsubscribed() {      return worker.isUnsubscribed();    }    @Override    public Subscription schedule(Action0 action, long delayTime, TimeUnit unit) {      if (threadPool != null) {        if (!threadPool.isQueueSpaceAvailable()) {          throw new RejectedExecutionException("Rejected command because thread-pool queueSize is at rejection threshold.");        }      }      return worker.schedule(new HystrixContexSchedulerAction(concurrencyStrategy, action), delayTime, unit);    }    @Override    public Subscription schedule(Action0 action) {      if (threadPool != null) {        if (!threadPool.isQueueSpaceAvailable()) {          throw new RejectedExecutionException("Rejected command because thread-pool queueSize is at rejection threshold.");        }      }      return worker.schedule(new HystrixContexSchedulerAction(concurrencyStrategy, action));    }  }

解决办法：

适当调大Hystrix线程队列参数动态水平扩容服务优化下游服务，减少服务响应时间