Open Bug 2030340 Opened 6 months ago Updated 5 months ago

[meta] [phase-2] fenix app links performance investigation work

Categories

(Firefox for Android :: Tooling, task)

All
Android
task

Tracking

()

People

(Reporter: kaya, Unassigned)

References

(Depends on 14 open bugs)

Details

(Keywords: meta, Whiteboard: [fxdroid][group6] )

Second phase meta bug to group together the app links performance work as we've stepped in Q2 2026 and completed most of the tickets in the original (phase 1) meta bug. The tickets linked to this meta include the ones we are carrying over from the original meta bug (lower priority tickets) and the ones that will be created from now on. We are closing the original meta ticket with this one replacing it.

Keywords: meta
Whiteboard: [fxdroid][group6]
Depends on: 1971437
Depends on: 1623925
Depends on: 2003090
Depends on: 2003950
Depends on: 1641129
Depends on: 2003953
Depends on: 2004793
Depends on: 2005263
Depends on: 1953033
Depends on: 1965899
Depends on: 1958039
Depends on: 1770989
Depends on: 1985895
Depends on: 1833222
Depends on: 1947487

The Bugbug bot thinks this bug should belong to the 'Firefox for Android::Tooling' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: General → Tooling

Wanted to point out that there was a recent regression in autofill metrics with bug 1979760, which is related to at least bug 1965899 and bug 1971437, and probably the app link meta in bug 1947487. I think it would be highly valuable if we were able to prioritize a structural solve to recording early metrics. There are a number of bugs I linked in bug 2034653 that have additional context. A few were discussed as we were investigating offline today:

  1. swapping these metrics to user lifetime in Glean, but this has undesirable downstream repercussions in associating the data very closely to users' profiles.
  2. continuing to investigating either an Android or Glean-defined cache which could be used until metrics were known to be stale
  3. deferring all the current startup metrics to a WorkManager Worker

Happy to chat through our experience with this issue last time if more context is useful!

You need to log in before you can comment on or make changes to this bug.