DanielLeens commented on issue #12268: URL: https://github.com/apache/seatunnel/issues/12268#issuecomment-5635774969
Thanks for putting the proposal in STIP form. It is now reviewable as a design, but it is not accepted or implemented. The current source makes the bounded-read premise the key unresolved point. `CoordinatorService.getJobInfo` reads retained history, then an active or pending `JobMaster`, then restores a logical DAG from persisted job information. `JobMaster.getJobDAGInfo` lazily builds the DAG, and the existing REST compatibility path can regenerate it when the cache misses. Those paths are not equivalent to a pure, cheap graph mapper. Before any implementation or mailing-list acceptance, please freeze a first-slice contract for: (1) selecting history versus an active job when an ID is reused; (2) whether this endpoint may ever restore or regenerate a DAG, and its explicit deadline/permit behavior if it can; (3) a topology-only field whitelist that never serializes environment options; and (4) bounded response size, retention states, and deterministic 404/409/503 outcomes. The contract tests should prove that a cache miss cannot silently trigger connector loading or unbounded reconstruction on a REST or scheduling thread. Keep this separate from #12207: no OpenLineage receiver, column lineage, or new persistent lineage store belongs in this first proposal. -- 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]
