It looks like a failure on the coqbot server led GitLab to disable the "job" webhook on Tuesday (2022-12-06) at 6.52pm GMT (7.52pm Paris time). Fortunately, the "pipeline" webhook is separate and was therefore not disabled. But this means that no direct job link was posted in the meantime, only the overall pipeline success or failure.
Their doc says (https://docs.gitlab.com/ee/user/project/integrations/webhooks.html#troubleshoot-webhooks) that just test-running the webhook again is supposed to re-enable it if it succeeds. I did that, so we'll see if the webhook is indeed re-enabled.
Let me know if you see this happening again (look for the banners like this on GitLab):
Webhook disabled: A webhook in this project was automatically disabled after being retried multiple times.
Is this weird bench failure related to the webhook issue? https://github.com/coq/coq/pull/16941/checks?check_run_id=9967727036
no, that's the bot not handling it when the bench dies badly (without generating the files in timings/)
Last updated: Feb 05 2023 at 22:03 UTC