[
https://issues.apache.org/jira/browse/KAFKA-12817?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17348619#comment-17348619
]
A. Sophie Blee-Goldman commented on KAFKA-12817:
------------------------------------------------
Thanks for filing this Bruno. I still have my own personal reservations about
this, but I'm happy to discuss it further if/when a KIP is raised
> Make Task ID an Implementation Detail
> -------------------------------------
>
> Key: KAFKA-12817
> URL: https://issues.apache.org/jira/browse/KAFKA-12817
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Bruno Cadonna
> Priority: Minor
> Labels: needs-kip
>
> Each task has a task ID that is used to identify tasks within Kafka Streams.
> The task ID is composed of the subtopology ID it executes and the number of
> the partitions the task reads its input data from. This naming scheme is
> rather an implementation detail and it is not something user should need to
> rely on to get metadata of a task. However, the task ID in this form is used
> to tag metrics, in log files, and its representation in code, the {{TaskId}}
> class is part of the public API.
> This ticket proposes to make the task ID really an implementation detail by:
> * removing {{TaskId}} from the public API
> * use the subtopology ID and the partition numbers in logs and metrics
> instead of the task ID
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
