[
https://issues.apache.org/jira/browse/HBASE-27268?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Duo Zhang resolved HBASE-27268.
-------------------------------
Fix Version/s: 2.5.0
3.0.0-alpha-4
2.4.14
Hadoop Flags: Reviewed
Resolution: Fixed
Pushed to branch-2.4+.
Thanks [~zhuyaogai] for contributing!
> In trace log mode, the client does not print callId/startTime and the server
> does not print receiveTime
> -------------------------------------------------------------------------------------------------------
>
> Key: HBASE-27268
> URL: https://issues.apache.org/jira/browse/HBASE-27268
> Project: HBase
> Issue Type: Improvement
> Components: tracing
> Affects Versions: 2.5.0, 3.0.0-alpha-3
> Reporter: zhuyaogai
> Assignee: zhuyaogai
> Priority: Minor
> Fix For: 2.5.0, 3.0.0-alpha-4, 2.4.14
>
>
> *Question:*
> In some scenarios in the production environment, users find that the request
> latancy is high on the client side, but it is not high on the server side
> (all through the internal monitor system). In this inconsistent scenario, we
> want to know if there is a problem with the client or the current network.
> Therefore, we want to know the time when the server receives the request.
> However, in Trace mode, no relevant information is printed.
>
> A possible solution is that the client prints callId (so we can associate the
> corresponding request on the server) and startTime about the request，and the
> server prints receiveTime about the request. In this way, we can find out at
> what stage the problem occurs.
>
> The client can even wrap the startTime in the request, and the server can
> directly collect the time at the network level (netty framework process time
> and network transmission time and so on) which make it easy to monitor some
> network level metrics about the client requests.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
