I have verified on re-dash that analzye failures is finding only 8 linux64 pgo tasks as high value, but for some reason we are running 100% of them every commit. I don't see any linux64-pgo/opt tasks as low value from the api: https://treeherder.mozilla.org/api/project/mozilla-inbound/seta/job-priorities/?build_system_type=taskcluster&format=json I suspect we have code that is ignoring pgo or not handling it properly, possibly we can look for how 'asan' is handled and make that work.
Aren't we running 100% of all tier-1 Taskcluster tests, because we shut off using SETA for them until we had a tier-1 supported way to backfill them?
we do have methods for backfilling now, although last night there was an issue with the credentials for taskcluster/treeherder conflicting.
we should be emitting low-value jobs like test-linux64-pgo/opt-*. I don't see us specifically blocking pgo in the decision task or in taskcluster- I assume we can figure this out with a bit more searching.