Public Repo Stale Check workflow incorrectly closes PRs right after adding stale label instead of after 3 days

[rakleed]

Code of Conduct

• I have read and agree to the GitHub Docs project's Code of Conduct

What article on docs.github.com is affected?

https://github.com/github/docs/blob/main/.github/workflows/triage-stale-check.yml

What changes are you suggesting?

You can see it, for example, here: #35028 (comment) — the label stale is added, and then immediately the PR closes. Although, according to the description and judging by the code, this should happen 3 days after that:

docs/.github/workflows/triage-stale-check.yml

Lines 29 to 31 in 08e8e07

	stale-pr-message: 'A stale label has been added to this pull request because it has been open 7 days with no activity. To keep this PR open, add a comment or push a commit within 3 days.'
	days-before-pr-stale: 7
	days-before-pr-close: 3

Additional information

No response

[subatoi]

Hi @rakleed and thanks for opening this issue—in that particular PR's case, I suspect it may have something to do with the fact that the PR was marked as stale multiple times. If you can show me an example of a PR where that hasn't been the case, please feel free to (even if this issue is closed).

Either way, I'm going to let our engineering team know about this and ask them to triage it internally in case any action is required. Since we regularly monitor this repo manually—including those issues that have gone stale—there shouldn't be any impact on contributors and I'm going to close this out, but as I said, feel free to add other example(s) latterly.

Thanks again for your interest in the GitHub docs!

[rakleed]

@subatoi when this happened for the first time in this PR, the PR was also closed immediately after the assignment of the stale label: #35028 (comment).

Here is another example where the stale label was assigned only once, but the PR was also immediately closed: #35829 (comment).

[subatoi]

@rakleed OK, thank you for pointing those out—I'll go ahead and create an issue for our engineering team, but we'll need to handle that internally to get to the bottom of it. As I said, fortunately there's no actual impact for contributors. Thanks again!