1875263 - 17.83% cnn-ampstories LastVisualChange (Android) regression on Tue January 9 2024

Status: RESOLVED WORKSFORME

(Testing :: General, defect, P1)

Description

[Andra Esanu (needinfo me)]

Perfherder has detected a browsertime performance regression from push d4ff02b92c066b55eaedf705b0bec7eab81178b7. 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)	Performance Profiles
18%	cnn-ampstories LastVisualChange	android-hw-a51-11-0-aarch64-shippable-qr	warm webrender	543.43 -> 640.33	Before/After

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)	Performance Profiles
39%	cnn-ampstories loadtime	android-hw-a51-11-0-aarch64-shippable-qr	cold webrender	1,234.22 -> 750.22	Before/After
31%	cnn-ampstories loadtime	android-hw-a51-11-0-aarch64-shippable-qr	cold webrender	1,282.29 -> 888.81	Before/After
30%	cnn fcp	android-hw-a51-11-0-aarch64-shippable-qr	cold webrender	2,170.80 -> 1,528.33	Before/After
23%	cnn-ampstories fcp	android-hw-a51-11-0-aarch64-shippable-qr	cold webrender	1,142.31 -> 884.79	Before/After
22%	cnn-ampstories SpeedIndex	android-hw-a51-11-0-aarch64-shippable-qr	cold webrender	1,326.95 -> 1,031.68	Before/After
...	...	...	...	...	...
4%	cnn-ampstories loadtime	android-hw-a51-11-0-aarch64-shippable-qr	warm webrender	295.36 -> 283.10	Before/After

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 patch(es) may 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.

You can run these tests on try with ./mach try perf --alert 41060

For more information on performance sheriffing please see our FAQ.

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1859911

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

13 improvements most of them large and 1 regression- all from changing a manifest file format? I relooked at my changes and there are no prefs or other factors that might be up in question about what to measure. Looking at the graphs, retriggers, and backfills this looks to be related to my change.

I see wikipedia was run on a few related changesets, but it didn't have fluctuating values? Given the nature of changes (to the harness, i.e. infra/tooling), I would have expected other webpages to show similar regressions, at least from the same manifest.

Assuming this is related, then I can only think of the toml library being used while loading the manifest on the host machine (not the phone) as a problem, maybe the test harness is using more memory? If that was the case, then this would show up on more than 1 page.

:aesanu, what is the process for marking this as a tooling or infra related issue? Are there other common steps I should take?

Comment 3

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1859911

Comment 4

[Andra Esanu (needinfo me)]

hey :jmaher, I will mark the alert as being infra. Thank you for looking into this.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1859911

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

The severity field is not set for this bug.
:jmaher, could you have a look please?

For more information, please visit BugBot documentation.

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

aesanu, did you mark this as infra? should this bug be closed?

Comment 8

[Andra Esanu (needinfo me)]

Yes, the alert is set to infra. The bug can be closed.