DanielLeens commented on issue #12750: URL: https://github.com/apache/seatunnel/issues/12750#issuecomment-6145194685
Yes — this is a useful follow-up, provided it remains a read-only evidence feature rather than a root-cause claim. Please wait until #12631 and #12688 are merged, then rebase the work on current `dev`; do not submit their pending changes as part of this PR. The proposed contract is a sound starting point: match only completed buckets in the two endpoint windows, preserve individual queue IDs and target vertices, report the measured window and sample count, and call unavailable, disabled, partial, or insufficient data a coverage gap rather than healthy behavior. Keep source-read and sink-write observations explicitly non-causal, and report prepare/commit pressure separately. For the first PR, keep the threshold rules fixture-driven and cover transient versus sustained pressure, missing or conflicting buckets, disabled observability, multiple queues sharing a target, idle upstream, and commit-only sink time. The CLI should recommend investigation of a candidate, never assert which component caused the slowdown. Please link the PR here once the dependencies have merged. -- 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]
