meta: re-enable PR stalebot

[avivkeller]

I propose that this stalebot be re-enabled, as there are hundreds(?) of stale PRs that are over a year old that could benefit from being closed.

[nodejs-github-bot]

Review requested:

• @nodejs/actions

[RafaelGSS]

I'm concerned that this will cause a lot of activity in our GitHub notification box and lead to important PRs being overlooked.

[avivkeller]

I'm concerned that this will cause a lot of activity in our GitHub notification box and lead to important PRs being overlooked.

While that was a concern with this in the past, @mhdawson addressed (AFAICT) by having this only apply to PRs older than a year. I'm happy to adjust that date, but I do think that a stalebot (in any form) is a good thing.

If one year is two soon, we can always start at an arbitrary value (say 3 years or something), and gradually lower the threshold to reduce noise.

[RafaelGSS]

Which value do you think this would bring in case of PRs?

I can see the value for issues, as they are normally a source of state-of-art of a project, but pull requests don't seem to influence anything at all.

[avivkeller]

Which value do you think this would bring in case of PRs?

Implementing a PR stalebot can help prevent accidental "cookie-licking" (https://devblogs.microsoft.com/oldnewthing/20091201-00/?p=15843). By automatically closing outdated and inactive PRs, it creates an opportunity for new or existing contributors to take on the work themselves, rather than waiting for an old, stale PR to be merged.

I can see the value for issues, as they are normally a source of state-of-art of a project, but pull requests don't seem to influence anything at all.

#54672

[avivkeller]

I can also, along with the above, lower the op count to ~30 (which is like 5-10 prs, if that), which will only close a few PRs for notification management. It can be increased as needed.