Uhh, what are you doing that relies on the naming being a certain format? There's no guarantee the PR can get the name that's generated unfortunately :/

Sometimes it will collide with another PR environment

In our Notion cards, we infer the PR environment link from the PR number. This way, when QA picks up the card, they can just click the link in the card that opens up the PR environment. We've been doing this for a while and haven't had any issues until today.

If you close and re-open a PR that we haven't fully cleaned up, this can happen. Were these re-opened PRs?

Yes, this is exactly what happened! I closed and reopened a PR following the issues we had with volume migrations last night. I just created a new PR and it looks like it's working as expected. Thanks for the help!

Of course! Behind the scenes, we go can delete the PR environment. That can take up to 1-2 minutes depending on containers, volumes, etc.

While we do this, we don't change the environment name until that environments fully cleaned up, in case something fails to delete.

And, just so you know this only happens if you close and re-open a PR in quick succession.

We're trying to make this faster. BTW looks like the "chemcloud-pr-1432" and "chemcloud-pr-1434" are now unused

So, if you close and re-open that PR (again) you should get the "right names"

Fantastic, thank you for the explanation. Very useful to know!

We're talking about this internally to make it so that this doesn't happen. Names are unique, so we add a hash when the delete is completed.

We can add this ahead of time, and if it fails, we can just show the generated hash or something. The DX is a bit odd so we'll have to work on it tho.

For now, if you close stuff, wait a 1-2 minutes before reopening and you should be good

Makes sense! I'll keep this in mind for next time this happens, but I don't foresee this being an issue again, as it was really just because of the other issue that I closed and opened the PR. Thank you again!