1744038 - 47.04 - 22.25% a11yr dhtml.html / a11yr dhtml.html (Linux, OSX) regression on Sat November 27 2021

Status: RESOLVED INVALID

(Core :: DOM: Navigation, defect)

Description

[Acasandrei Beatrice (needinfo me)]

Perfherder has detected a talos performance regression from push 56a1fce20849e569b75856301c801f89a0d6e0b9. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio	Test	Platform	Options	Absolute values (old vs new)
47%	a11yr dhtml.html	linux1804-64-shippable-qr	e10s fission stylo webrender	380.96 -> 560.16
22%	a11yr dhtml.html	macosx1015-64-shippable-qr	e10s fission stylo webrender	319.62 -> 390.73

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
3%	tresize	linux1804-64-shippable-qr	e10s fission stylo webrender	21.63 -> 21.00

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.

Comment 1

[Acasandrei Beatrice (needinfo me)]

I'm not sure which one of these bugs (Bug 1741132, Bug 1740517 or Bug 1740516) caused the above regression. Can you provide more information? Thanks.

Comment 2

[Peter Van der Beken [:peterv]]

(In reply to Acasandrei Beatrice from comment #1)

I'm not sure which one of these bugs (Bug 1741132, Bug 1740517 or Bug 1740516) caused the above regression. Can you provide more information?

Bug 1740516 and bug 1740517 only affect loads from session history (so going backwards in session history). Bug 1741132 makes us keep less pages in the BFCache, which also only affects loads from session history. Unless I'm missing something these tests don't navigate in session history, so it seems highly unlikely that these caused the regression.

Alexander, am I right in thinking that the a11y dhtml.html test doesn't navigate in session history?

Comment 3

[James Teh [:Jamie]]

Eitan, do you know anything about the a11y talos tests? These were before my time and I don't know anything about them. :(

Comment 4

[James Teh [:Jamie]]

The baseline for these tests was changed as a result of Cache the World in bug 1801330.