unhappy user ID 104802998

Stream: general

[Damiano Testa]

Dear All,

in #13483 I keep getting a failed issue with

 .github/workflows/merge_conflicts.yml #46440

I tried merging master and changing the file slightly to produce a new commit, but the error stays there. Note that #14040 which is based on top of #13483 does not have this issue.

What am I doing wrong?

Thanks!

[Eric Wieser]

You're not doing anything wrong, this is a new problem

[Eric Wieser]

The issue will appear or disappear depending on whether it's a busy time of day when you push it

[Damiano Testa]

Ok, thanks! I was then unlucky twice, I'll try my luck again.

[Damiano Testa]

I've hit this error for yet another time: this is looking like too much luck, no?

[Eric Wieser]

The rate limiting is an hour long window

[Eric Wieser]

I don't know if that's a rolling window or resets on the hour

[Damiano Testa]

Ok, I tried 3 hours ago, 2 hours ago, 43 minutes ago and 6 minutes ago.

I'll wait another hour!

[Damiano Testa]

It seems that if the PR builds sufficiently quickly, the unhappy 104802998 does not cancel the build, so you still get oleans.

I may not play the game of "catch 104802998 while they are happy"!

[Bryan Gin-ge Chen]

From what I recall, a failing merge_conflicts job shouldn't be able to prevent the creation of oleans (and a failing status for that job shouldn't block bors either). It's a bit annoying but I think you can proceed as if that particular red x isn't there.

[Damiano Testa]

Thanks Bryan! Indeed, most of the red crosses that I see actually have oleans and passed linting and tests.