Perfherder has detected a browsertime performance regression from push e6cd9ff9eca2b2b3c45abde7fd8b62e79cf1aa38. 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.

For more information on performance sheriffing please see our FAQ.

The test seemed bimodal before the patch, fwiw.

(In reply to Ionuț Goldan [:igoldan] from comment #0)

Perfherder has detected a browsertime performance regression from push e6cd9ff9eca2b2b3c45abde7fd8b62e79cf1aa38. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
19%	amazon	loadtime	macosx1014-64-shippable-qr	cold nocondprof webrender	1,739.04 -> 2,072.04

Curious that we saw both a 20% regression in Amazon.com page load time and a 5% improvement in Amazon.com page load's visual metrics (bug 1690842 comment 6) after non-native theming was enabled on macOS.

needinfo'ing Jesup for awareness, in case he has some insights.

Looking at the graph over the last year, it looks like enabling non-native theming may actually have stabilized this test and possibly made it more meaningful. We used to fluctuate quite heavily. Although we have stabilized at the high end, we may be able to measure improvements more accurately going forward.

(In reply to Stephen A Pohl [:spohl] from comment #3)

Looking at the graph over the last year, it looks like enabling non-native theming may actually have stabilized this test and possibly made it more meaningful. We used to fluctuate quite heavily. Although we have stabilized at the high end, we may be able to measure improvements more accurately going forward.

Like Stephen said, non-native theming seems to have made this noisy, bimodal test stable with a slightly better worst-case than before:

https://treeherder.mozilla.org/perfherder/graphs?highlightAlerts=1&highlightChangelogData=1&series=autoland,2899640,1,13&series=mozilla-central,3424276,1,13&timerange=31536000

btw, this test has been moved from an old Mac Mini running macOS 10.14 to a newer, faster Mac Mini running macOS 10.15. The test results are still stable, but faster on the newer hardware.

I think we can resolve this bug as WONTFIX.

Resolving per comment 4.