E.g.

https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightCommonAlerts=0&series=mozilla-central,3737968,1,13&series=mozilla-central,3737968,1,13&series=autoland,3911355,1,13&timerange=5184000&zoom=1680453491826,1680765194016,152.9664831386612,264.44110804093174

https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightCommonAlerts=0&series=mozilla-central,3738034,1,13&series=mozilla-central,3738034,1,13&series=autoland,3911421,1,13&timerange=5184000&zoom=1680573415534,1680755348102,26.870918602135248,41.417100822011044

suspected: bug 1819722

feel free to WONTFIX.

Set release status flags based on info from the regressing bug 1819722

:dthayer, since you are the author of the regressor, bug 1819722, could you take a look? Also, could you set the severity field?

For more information, please visit auto_nag documentation.

https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightCommonAlerts=0&selected=3911357,1675282111&series=mozilla-central,3737970,1,13&series=mozilla-central,3737970,1,13&series=autoland,3911357,1,13&timerange=5184000&zoom=1680590754516,1680723586775,153.88805372582715,329.6849944411657

Sorry, yeah I think the conclusion is going to end up being WONTFIX here, but I'm going to assign myself just to verify it's the same issue as Kraken, and also try out tuning the heuristic that's hitting this to see if we can make our benchmarks any happier. In any case it's certainly not reason enough to revert the change.

Set release status flags based on info from the regressing bug 1819722

Hi Doug, any updates on this investigation?

(In reply to Ryan VanderMeulen [:RyanVM] from comment #5)

Hi Doug, any updates on this investigation?

Ah! Yes, sorry. It is indeed just sensitive to the inlining entry threshold heuristic. However, Speedometer and other more complex and real-worldy benchmarks are sensitive to this heuristic in the other direction. Since we care about those more, and JetStream 2 subtests seem to perform very differently if you just breathe on them wrong anyway, I think we should just WONTFIX this. Going to go ahead and do that, but feel free to undo it if you disagree.