rusackas commented on PR #44538: URL: https://github.com/apache/superset/pull/44538#issuecomment-5803955361
Fair, and looking at the numbers I think you're right. Entering the queue doesn't reuse the PR's existing CI, GitHub only runs a workflow against the merge-group ref if it's wired for that event, so every queued entry gets a full second run of the required-check matrix (Playwright, Cypress, the DB-flavor suites) on top of its normal PR-time run. With `min_entries_to_merge: 1` and how often master merges right now, most of those land as a batch of one, so there's not much batching savings in practice, closer to +100% required-check compute per merge than an occasional cost. Against that, the incident this was meant to prevent (#44465 landing 18 minutes after #44467 regenerated `messages.pot`, since #44465's branch predated the new drift check entirely) looks like a one-time bootstrapping gap rather than something that happens often. Holding this for now, don't think the tradeoff clears the bar at our scale. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
