junaiddshaukat commented on issue #39135:
URL: https://github.com/apache/beam/issues/39135#issuecomment-4825597293
> could you please help me with the above quesitions and review the pr
Hi @boy397 — thanks for the interest! This is a GSoC 2026 sub-issue for the
Kafka Streams runner that I'm implementing as the GSoC contributor (part of the
project skeleton under #18479), so it's already assigned to me and the
implementation is up in #39141 — that answers both your questions:
1. Watermark fan-out: there's a broadcast partitioner on the repartition sink
that sends the watermark report to all partitions; GroupByKeyProcessor
feeds
those into the existing WatermarkManager and fires the groups once when
the
watermark reaches TIMESTAMP_MAX. So no change to WatermarkManager itself —
the broadcast happens at the sink + the processor handles propagation.
2. Repartition topic: for now it's expected to pre-exist (same assumption as
the
Impulse bootstrap topic); runner-side auto-creation via AdminClient is a
separate follow-up, not part of this issue.
Since this one is inside my GSoC scope, my mentor has asked that these
runner sub-issues stay with the GSoC work rather than separate PRs, so I'd ask
you to hold off on #39136 here — but please don't let that put you off Beam!
There's plenty to dig into; the "good first issue" / "newbie" labels are a
great place to start. Thanks for understanding 🙂
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
