Let's repurpose this bug, expanding on Devan's comments below:

<dgoodwin> This caused many test regressions in component readiness, not sure what it was but it reasolved 4 days ago after a clear pattern of permafailing a whole bunch of tests.

If anyone could add details to this bug that would be nice, but I'm going to immediately close the bug, we just need something to traige the board.

<lmeyer> This happened again three days ago, and it was clearly not a permafailing test; it was all of the jobs in an aggregated run failing the same way, against a single payload.

I don't want to link all the tests that regressed, because most likely they had nothing to do with whatever actually failed. But have a look at a two-week report of the regressed test this bug originally reported and check two jobs that failed, the first on 5/22, and the second on 5/27. This is a pattern.