DanielLeens commented on issue #10997: URL: https://github.com/apache/seatunnel/issues/10997#issuecomment-5098260301
Thanks for the controlled validation and for splitting the findings. I agree with the current classification: the original #10997 remains a real Zeta control-plane / completion-path scalability issue, but the latest evidence narrows the immediate patch to isolating metrics reporting from Hazelcast generic-operation threads rather than proving a deterministic before/after reproduction of the permanent `RUNNING` hang. PR #11564 is now the right review path for the `ReportMetricsOperation` part. The important boundary is that it keeps the metrics payload, serializer, storage, and `SeaTunnelServer.updateMetrics()` behavior unchanged, while moving the expensive work off the generic-operation thread. The separate Worker-restart checkpoint recovery NPE is correctly tracked in #11563 and #11565. I removed the stale `help wanted` label because there is now an active same-root-cause PR. Next, let us use #11564 current-head Build and maintainer review as the implementation gate, and keep this issue open until we know whether that patch is sufficient or whether final-checkpoint retry / timeout handling still needs a follow-up. -- 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]
