[
https://issues.apache.org/jira/browse/TUBEMQ-313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17176047#comment-17176047
]
Guocheng Zhang commented on TUBEMQ-313:
---------------------------------------
Well, this is indeed a problem. When the signaling layer and the data layer are
mixed together, and the data layer requests are very large, each request
processing time is very long, this is a problem, let us see if there is any
good way to solve.
> Broker heartbeat performance issues
> ------------------------------------
>
> Key: TUBEMQ-313
> URL: https://issues.apache.org/jira/browse/TUBEMQ-313
> Project: Apache TubeMQ
> Issue Type: Improvement
> Components: Broker
> Reporter: charleli
> Priority: Major
>
> Broker heartbeat performance issues
> At present, Broker needs to report heartbeats regularly. In the case of large
> batch requests, when the CPU performance of the client and server is very
> high, the heartbeat report is not timely, and there is a high probability
> that the heartbeat report fails, causing the client to disconnect.
> At the optimization point, the data packet request of the client can be
> regarded as the heartbeat data.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
