[BUG] metric timer will trigger integer overflow .

[FMX]

What is the bug?

RssWorker metric system has a bug that will cause an integer overflow when runs long enough.

How to reproduce the bug?

Steps to reproduce the bug.

Could you share logs or screenshots?

If applicable, add logs/screenshots to help explain your problem.

22/01/06 10:24:53,820 ERROR [push-server-6-10] TransportRequestHandler: Error while invoking RpcHandler#receive() on PushData PushData{requestId=702699486, mode=1, shuffleKey=application_1640695558204_38730_1-2, partitionUniqueId=151-0, body size=15066}
java.lang.ArrayIndexOutOfBoundsException: -4095
at com.aliyun.emr.rss.server.common.metrics.ResettableSlidingWindowReservoir.update(ResettableSlidingWindowReservoir.scala:35)
at com.codahale.metrics.Histogram.update(Histogram.java:39)
at com.codahale.metrics.Timer.update(Timer.java:164)
at com.codahale.metrics.Timer.update(Timer.java:86)
at com.aliyun.emr.rss.server.common.metrics.source.AbstractSource.doStopTimer(AbstractSource.scala:148)
at com.aliyun.emr.rss.server.common.metrics.source.AbstractSource.stopTimer(AbstractSource.scala:131)
at com.aliyun.emr.rss.service.deploy.worker.Worker$$anon$7.onSuccess(Worker.scala:587)
at com.aliyun.emr.rss.service.deploy.worker.Worker.handlePushData(Worker.scala:660)
at com.aliyun.emr.rss.service.deploy.worker.PushDataRpcHandler.receivePushData(PushDataRpcHandler.java:56)

22/01/06 10:25:31,123 WARN [nioEventLoopGroup-11-1] DefaultChannelPipeline: An exceptionCaught() event was fired, and it reached at the tail of the pipeline. It usually means the last handler in the pipeline did not handle the exception.
java.lang.NegativeArraySizeException
at com.aliyun.emr.rss.server.common.metrics.ResettableSlidingWindowReservoir.getSnapshot(ResettableSlidingWindowReservoir.scala:40)
at com.codahale.metrics.Histogram.getSnapshot(Histogram.java:54)
at com.codahale.metrics.Timer.getSnapshot(Timer.java:159)
at com.aliyun.emr.rss.server.common.metrics.source.AbstractSource.recordTimer(AbstractSource.scala:251)
at com.aliyun.emr.rss.server.common.metrics.source.AbstractSource$$anonfun$getMetrics$4.apply(AbstractSource.scala:282)

/cc @waitinfuture

/assign @FMX

[FMX]

I think it`s fixed.