1720072 - 100 - 35.55% tp5o_webext XRes / cpstartup content-process-startup + 4 more (Linux) regression on Tue July 6 2021

Status: RESOLVED FIXED

(Testing :: Talos, defect)

Description

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Perfherder has detected a talos performance regression from push 003db85b60d01fa3dde0f026a44719cee8b7c912. As author of one of the patches included in that push, we need your help to address this regression.

Improvements:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
100%	tp5o	XRes	linux1804-64-shippable-qr	e10s stylo webrender	198,823.46 -> 0.00
100%	tp5o	XRes	linux1804-64-shippable-qr	e10s stylo webrender-sw	6,486,677.36 -> 0.00
100%	tp5o_webext	XRes	linux1804-64-shippable-qr	e10s stylo webrender-sw	6,484,976.28 -> 0.00
100%	tp5o_webext	XRes	linux1804-64-shippable-qr	e10s stylo webrender	198,850.89 -> 0.00
39%	cpstartup	content-process-startup	linux1804-64-shippable-qr	e10s stylo webrender	100.33 -> 61.25
36%	cpstartup	content-process-startup	linux1804-64-shippable-qr	e10s stylo webrender-sw	96.33 -> 62.08

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

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

Looks like a bug in [the testing code here]: content processes will generally no longer show up in xrestop output because they're not using X, so that should continue the loop instead of throwing an exception. (They may still appear if WebGL is active when the sample is taken.)

It looks like that exception is caught here and discarded, which would explain why the Could not find PID message doesn't appear in the logs.

Note that the cpstartup changes are an improvement, and expected.

I'll take the bug; it should be simple.

Comment 2

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

Attachment: Bug 1720072 - Allow the XRes perf metric to tolerate content processes not using X11.

Comment 3

[Chris Peterson [:cpeterson]]

Should we uplift this fix to Beta 91? Semi-headless mode is riding the trains in Fx91, but I don't know if we monitor Talos performance in channels other than mozilla-central/Nightly.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/8fe683c45800
Allow the XRes perf metric to tolerate content processes not using X11. r=perftest-reviewers,AlexandruIonescu

Comment 5

[Dorel Luca [:dluca]]

https://hg.mozilla.org/mozilla-central/rev/8fe683c45800

Comment 6

[Chris Peterson [:cpeterson]]

Alexandru, should we uplift Jed's Talos cmanager_linux.py fix to Beta 91? Do we run Talos on Beta builds?

Comment 7

[Alexandru Ionescu (needinfo me) [:alexandrui]]

(In reply to Chris Peterson [:cpeterson] from comment #6)

Alexandru, should we uplift Jed's Talos cmanager_linux.py fix to Beta 91? Do we run Talos on Beta builds?

Seems like we do.

Comment 8

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

The patch landed in nightly and beta is affected.
:jld, is this bug important enough to require an uplift?
If not please set status_beta to wontfix.

For more information, please visit auto_nag documentation.

Comment 9

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

Comment on attachment 9230781 [details]
Bug 1720072 - Allow the XRes perf metric to tolerate content processes not using X11.

Beta/Release Uplift Approval Request

• User impact if declined: Missing performance metrics. (We track these on m-c and beta, but apparently not release or ESR.)
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce: Run Talos tp5o on Linux; check if the XRes metric is 0 (it shouldn't be).
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): NPOTB: affects only performance test code
• String changes made/needed: none

Comment 10

[Pascal Chevrel:pascalc (PTO until April 26)]

Comment on attachment 9230781 [details]
Bug 1720072 - Allow the XRes perf metric to tolerate content processes not using X11.

Approved for 91 beta 5, thanks.

Comment 11

[Pascal Chevrel:pascalc (PTO until April 26)]

https://hg.mozilla.org/releases/mozilla-beta/rev/fdc4c2fcc9f5