boomanaiden154 wrote: > Not to point fingers, but are we doing any tracking of test flakiness? I added an endpoint to the premerge advisor that should expose this info, but have never acted upon it. https://github.com/llvm/llvm-project/pull/224861