DanielLeens opened a new issue, #12126: URL: https://github.com/apache/seatunnel/issues/12126
### Code of Conduct - [x] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct) ### Search before asking - [x] I had searched in the [issues](https://github.com/apache/seatunnel/issues?q=is%3Aissue+label%3A%22bug%22) and found no similar issues. ### Describe the proposal This umbrella tracks the reliability-hardening work that remains after a source review of Zeta on `dev` commit `97d461bc0773399d632fd078735736ecd44f5f0b` (2026-09-05). Every referenced line was read at that commit with `git show`; no runtime fault injection or load test was run for this umbrella, and the sub-tasks are worded accordingly: confirmed code behavior is stated as fact, failure chains that have not been reproduced are labelled as hypotheses, and the rest are design improvements with their constraints spelled out. Scope: the admission path (`submitJob` -> `JobMaster.init()` -> pending queue -> `preApplyResources`), the coordinator and worker executors, terminal-state delivery, checkpoint scheduling, history listing, the `JobStatus` ordinal contract, and worker-side execution-state reconciliation after a cluster merge. Existing tracks that this umbrella relies on but does not duplicate: - #11437 (fixed-slot job stuck `PENDING/RESTORE`, worker slots leaked after active-master kill) with fix PR #11458. - #12010 (interrupted pending-job insertion acknowledged as success) with fix PR #12052. - #9807 (job reports `RUNNING` after leader kill while the sink stops receiving changes); no root cause is asserted here. - #10993 (submitted tasks remain `PENDING` with fixed slots and `WAIT`); root cause still open. - #10767 (differentiated retention of historical job runtime information). Already fixed on `dev` and verified present at the baseline: #10568 via #10567 (`TaskExecutionService.java:506-525`) and #10675 via #10692 (`CoordinatorService.java:1106-1163`). Deliberately not filed as sub-tasks: the dynamic-slot normalization of `job-schedule-strategy` to `REJECT` is documented behavior (`docs/en/engines/zeta/hybrid-cluster-deployment.md`), and the internal per-task metric dimensions in `TaskExecutionService.provideDynamicMetrics()` are not exported as Prometheus labels, so neither has a concrete defect to track; both are noted here only for completeness. The analysis was prepared with AI-assisted source review and then checked line by line against the pinned commit; an independent maintainer review of the draft was incorporated before filing. ### Task list Ordered by dependency: the two executor items must preserve lifecycle progress and terminal-fact delivery before any bounding is safe; the admission contract and queue ordering are designed together; the remaining items are independent. - [ ] #12117 - [ ] #12118 - [ ] #12119 - [ ] #12120 - [ ] #12121 - [ ] #12122 - [ ] #12123 - [ ] #12124 - [ ] #12125 ### Are you willing to submit PR? - [ ] Yes I am willing to submit a PR! -- 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]
