1790906 - Identify cause of temporary 262ms improvement to Firefox Focus start up & try to ship it to users

Status: RESOLVED INVALID

(Core :: Performance, task, P2)

Description

[Michael Comella (:mcomella) [NI reported issues only: ex-Mozilla]]

Attachment: Fenix dashboard screenshot

In our Nightly Firefox Focus cold VIEW start up test (i.e. measuring process launch to navigation start for an app link), we saw an improvement of 262ms on August 11 and a regression back to where we were on August 19. In Fenix Nightly, we noticed a similar trend: 121ms regression by August 11 undone by August 19.

Due to the similarity, we suspect the root cause is in platform code. Our geckoview_example cold ContentfulSpeedIndex test (i.e. the test above plus a page load) on amazon shows something similar: until August 9, we see slow outliers above 1800ms which stops until Aug 21, backing up our conclusion.

The goal of this bug is to identify the root cause of the improvement and determine if we can ship that improvement to users permanently.

I put this in Core::Performance because I'm not sure which component it should go into.

Performance recording (profile)

Profile URL: Haven't taken

System configuration:

• Fenix & focus tests are taken on my Moto G5
• geckoview_example tests ran on A51 devices in CI

Comment 1

[Michael Comella (:mcomella) [NI reported issues only: ex-Mozilla]]

Attachment: Focus dashboard screenshot

Comment 2

[Michael Comella (:mcomella) [NI reported issues only: ex-Mozilla]]

Attachment: geckoview_example dashboard screenshot

Comment 3

[Bas Schouten (:bas.schouten)]

Sadly the Performance Triage Calculator isn't particularly appropriate for determining an impact for this type of issue. But purely based on the significance of this improvement I'm going to rate this as high for now.

Comment 4

[Frank Doty [:fdoty]]

Next step is to run this is geckoview_example, as well as collecting profiles

Comment 5

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

The severity field for this bug is set to S3. However, the Performance Impact field flags this bug as having a high impact on the performance.
:dthayer, could you consider increasing the severity of this performance-impacting bug? Alternatively, if you think the performance impact is lower than previously assessed, could you request a re-triage from the performance team by setting the Performance Impact flag to ??

For more information, please visit BugBot documentation.

Comment 6

[Alex Thayer [:alexical] (she/her)]

I looked deeper into this a while back and it looked overwhelmingly likely that this was a red herring. I don't have my notes from then but I still think we should just close this.

Comment 7

[Frank Doty [:fdoty]]

Closing this bug based on comment 6 - if we find reasons to investigate this in the future, please open a new bug.