[ 
https://issues.apache.org/jira/browse/KAFKA-18807?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Sean Quah updated KAFKA-18807:
------------------------------
    Description: 
When events arrive at the event processor with less than 
1000/group.coordinator.threads ms between them, the thread idle time is rounded 
down to 0 because of integer division.

eg. when the group coordinator has 8 threads, and it receives >1k offset commit 
requests per second evenly split across shards, the thread idle time metric 
will be under-reported.

https://github.com/apache/kafka/blob/2353a7c5081b53402d5092afd255e2bdf7aa2d20/coordinator-common/src/main/java/org/apache/kafka/coordinator/common/runtime/MultiThreadedEventProcessor.java#L141

  was:
When events arrive at the event processor with less than 
1/group.coordinator.threads ms between them, the thread idle time is rounded 
down to 0 because of integer division.

eg. when the group coordinator has 8 threads, and it receives >1k offset commit 
requests per second evenly split across shards, the thread idle time metric 
will be under-reported.

https://github.com/apache/kafka/blob/2353a7c5081b53402d5092afd255e2bdf7aa2d20/coordinator-common/src/main/java/org/apache/kafka/coordinator/common/runtime/MultiThreadedEventProcessor.java#L141


> Group coordinator under-reports thread idle time
> ------------------------------------------------
>
>                 Key: KAFKA-18807
>                 URL: https://issues.apache.org/jira/browse/KAFKA-18807
>             Project: Kafka
>          Issue Type: Bug
>          Components: group-coordinator
>            Reporter: Sean Quah
>            Assignee: Sean Quah
>            Priority: Minor
>             Fix For: 4.0.0
>
>
> When events arrive at the event processor with less than 
> 1000/group.coordinator.threads ms between them, the thread idle time is 
> rounded down to 0 because of integer division.
> eg. when the group coordinator has 8 threads, and it receives >1k offset 
> commit requests per second evenly split across shards, the thread idle time 
> metric will be under-reported.
> https://github.com/apache/kafka/blob/2353a7c5081b53402d5092afd255e2bdf7aa2d20/coordinator-common/src/main/java/org/apache/kafka/coordinator/common/runtime/MultiThreadedEventProcessor.java#L141



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to