Perfherder has detected a browsertime performance regression from push d68d487f283810647232dd3b0c363a2f86586205. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the offending patch(es) will be backed out in accordance with our regression policy.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

For more information on performance sheriffing please see our FAQ.

Set release status flags based on info from the regressing bug 1717806

Hi Alexandru! The regression range you linked does not really match the regression bug. Which one is the correct one? Keeping the NI for Tom in case this is a shim regression.

Ah, there is a bug in the patch in bug 1717806. I'm not convinced that it's causing this, but I have a patch I'll land over the next couple of days so we can find out.

I'm sorry, I updated the description of the bug.

https://hg.mozilla.org/mozilla-central/rev/a752b5b8b54f

:twisniewski just a reminder that Monday, June 27th is merge day for 103 to beta. If there is further work needed on this bug?

:dmeehan: There's nothing more that I can personally do, from what I can tell. If this hasn't been resolved with the patch that I landed a week ago, then it wasn't related to my patches.

(In reply to Thomas Wisniewski [:twisniewski] from comment #9)

:dmeehan: There's nothing more that I can personally do, from what I can tell. If this hasn't been resolved with the patch that I landed a week ago, then it wasn't related to my patches.

Thanks for looking,

:alexandru do you have another regression range?

Indeed, seems like few days later a fix landed but the graaph is still noisier than before.

Set release status flags based on info from the regressing bug 1717806

:twisniewski looks like the patch introduced by alexandrui may not have been the cause for the regression. Is there another regression range to review for a potential fix?

Unfortunately I don't know, and looking at the parent and child patches near it on autoland isn't giving me any obvious clues, either.

Could it be related to bug 1741787? It seems the noise started increasing around the time it landed.

Maybe Eng_Esther is good to needinfo here too? Eng_Esther, do you think this might be regressed by the patches in bug 1741787 ?

Based on this graph: https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&highlightCommonAlerts=0&selected=3777014,1540089873&series=autoland,3777014,1,13&timerange=5184000&zoom=1654692974710,1654705016345,994.4392036697246,1151.717360932369

It's bug 1741787 that caused this. :alexandrui, could you double-check?
Here's the regression bug for that one: https://bugzilla.mozilla.org/show_bug.cgi?id=1773629

According to the graph, there's a valid regression here. I retriggered the revision to try reveal the noise and confirm it's just the instability of the test.

The retriggers confirmed the regression. I'm going to backout the patch and push to try see if there are any changes.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=05f0094ae3be8eec850507133f6de4786bf4c231

:alexandrui, in case you ultimately do want to back the patch out, before you do please check with me, as it would be good to make sure we don't have conflicts with other incoming patches that are going to change the version number in the related manifest.json.

The try push reveals that the patch caused a regression, but immediately after that it became progresively noisier so I'm not sure what we can do about it. I'll try backing it out on the latest autoland, and see the results.
Won't back it out without notice.

:alexandrui, try doing a test with this patch applied: https://phabricator.services.mozilla.com/D151757
It should reduce a lot of the noise.

that patch brought us back to the numbers before the regression.

Based on comment #18 and comment #24, can you mark this bug as regressed by bug 1741787 and depends on bug 1779468.