[
https://issues.apache.org/jira/browse/KAFKA-20888?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18144744#comment-18144744
]
sanghyeok An commented on KAFKA-20888:
--------------------------------------
Related PR https://github.com/apache/kafka/pull/24234
> Expose Streams group member rack IDs in kafka-streams-groups.sh
> ---------------------------------------------------------------
>
> Key: KAFKA-20888
> URL: https://issues.apache.org/jira/browse/KAFKA-20888
> Project: Kafka
> Issue Type: Improvement
> Components: streams, tools
> Reporter: sanghyeok An
> Assignee: sanghyeok An
> Priority: Minor
> Labels: kip
>
> KIP-1370:
> [https://cwiki.apache.org/confluence/spaces/KAFKA/pages/446071011/KIP-1370+Include+Instance+and+Rack+IDs+in+kafka-streams-groups.sh+Member+Output]
>
> The Streams Rebalance Protocol tracks each member’s optional rack ID and
> exposes it through Admin#describeStreamsGroups as
> StreamsGroupMemberDescription#rackId().
> However, kafka-streams-groups.sh --describe --members does not currently
> display this information. Even with --verbose, the output includes member and
> assignment metadata such as process ID, member epoch, topology epoch, and
> assignments, but omits the rack ID.
> h3. Motivation
> Displaying rack IDs would help operators:
> * Verify that client.rack is propagated correctly.
> * Diagnose task-placement behavior when using a custom rack-aware task
> assignor.
> * Identify unexpected rack metadata changes that can trigger group epoch and
> assignment recomputation.
> * Inspect rack information without writing code against the Admin API.
> Because rack ID is optional and not needed for routine group inspection, it
> should be included only in verbose member output to avoid expanding the
> default table unnecessarily.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)