88fantasy commented on PR #12209: URL: https://github.com/apache/seatunnel/pull/12209#issuecomment-6438258184
Thanks @DanielLeens for the review. I have pushed the updated head `c8cdfaae7`: - Synced the latest shared lineage changes from #12208 and merged `dev@c875a87bd`. The conflict in `StateTransitionCleanupTest` was limited to imports; both the upstream tests and the lineage tests are preserved. GitHub now reports the PR as mergeable. - Coalesced pending heartbeats per job. A START or terminal event can replace a queued heartbeat-only job when the shared queue is full, while events for the same job remain ordered. - Limited the shared queue to 64 waiting jobs with two delivery threads. When the queue is full and no heartbeat-only job can be replaced, the new lifecycle event is dropped and the job ID and event type are logged. - Avoided a duplicate START after a failed savepoint start, while retaining the first START for a restored plan. Also updated the English and Chinese delivery documentation and the requested import. **Delivery limitation:** this remains best-effort. A terminal event can still be dropped when all 64 waiting slots contain lifecycle events, on shutdown, or after HTTP retries are exhausted. Therefore, this does not guarantee terminal delivery under sustained saturation; I would appreciate your view on whether this bounded behavior is acceptable for this PR. Local validation on this head with JDK 11: - `./mvnw spotless:apply` and full-reactor `./mvnw -q -DskipTests verify` passed. - 122 focused unit tests, 2 local engine IT tests, and 4 Markdown tests passed, with no failures or skips. The exact-head [fork Build](https://github.com/88fantasy/seatunnel/actions/runs/40479811555) is queued; local results do not establish CI success. Could you please re-review when convenient? -- 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]
