ci: restarting individual buildkite jobs doesn't update failed check status

#152
Opened by sterni at 2021-10-01T10·47+00

Sometimes we have “flaky” failures in the buildkite pipeline, the most common one is if whitby runs out of nixbld users. Restarting the failed jobs individually then doesn't update the overall failure status (tracked via buildkite metadata).

A workaround is to instead restart the whole pipeline, which requires rebuilding the pipeline.

  1. sterni changed the subject of this issue from "ci: restarting individual biuldkite jobs doesn't update failed check status" to "ci: restarting individual buildkite jobs doesn't update failed check status" at 2021-10-01T10·47+00
  2. Duplicate of b/152.

    sterni at 2021-10-01T10·49+00

  3. err, b/65.

    sterni at 2021-10-01T10·49+00

  4. sterni closed this issue at 2021-10-01T10·49+00