Closed Bug 1621948 Opened 5 years ago Closed 5 years ago

Re-enable (and fix?) the NUMBER_OF_PROFILES probe

Categories

(Toolkit :: Startup and Profile System, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla76
Tracking Status
firefox76 --- fixed

People

(Reporter: RT, Assigned: mossop)

References

Details

Attachments

(3 files)

NUMBER_OF_PROFILES (Number of named browser profiles for the current user, as reported by the profile service at startup) is disabled since 69 and it would be a useful probe to use in order to track multi profile generating installations so that such installations could be ignored from retention analysis.

Attached file Data collection request. (deleted) —
Attachment #9132868 - Flags: data-review?(chutten)

As part of this request I would like to check if there are issues with the probe.
https://sql.telemetry.mozilla.org/queries/69109/source#174376 shows the following weird things:

  • Installations with 'null' number of profiles - about a third of all reports
  • Installations with 1 profile - about 2 thirds of all reports (makes sense)
  • Installations with 2 profiles - only 1
  • Nothing else

Could there be an issue with the probe? I'd expect at least a share of installs to have more than 1 profile.
See also https://bugzilla.mozilla.org/show_bug.cgi?id=1296606#c27 that looked at why 11% of installs had 0 profiles

I think there may be an error in that query. I'm not sure what type histogram_parent_number_of_profiles is, but it should be a STRUCT, so the GROUP BY should fail. Here's a query using telemetry.main instead of main_summary which looks at the sum of the histogram:

https://sql.telemetry.mozilla.org/queries/69110/source#174377

The values are about what I'd expect given that most users reporting on March 3 are going to be running versions of Firefox >= 70... except for those reporting 0, those cheeky chums.

Comment on attachment 9132868 [details] Data collection request. Renewals now have their own dedicated (short!) form to follow. Could you give it a whirl for us? https://github.com/mozilla/data-review/blob/master/renewal_request.md
Attachment #9132868 - Flags: data-review?(chutten)

The priority flag is not set for this bug.
:mossop, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(dtownsend)

Romain, do you still want this?

Flags: needinfo?(dtownsend) → needinfo?(rtestard)
Attached file renewal_request.md (deleted) —

Thanks for the pointer, here is the short form filled in and sorry for the delay!

Flags: needinfo?(rtestard)
Attachment #9137041 - Flags: data-review?(chutten)

Apologies for the delay, a lot going on...
Regarding the request, thanks for the pointer, I attach the response.

(In reply to Chris H-C :chutten from comment #3)

I think there may be an error in that query. I'm not sure what type histogram_parent_number_of_profiles is, but it should be a STRUCT, so the GROUP BY should fail. Here's a query using telemetry.main instead of main_summary which looks at the sum of the histogram:

https://sql.telemetry.mozilla.org/queries/69110/source#174377

The values are about what I'd expect given that most users reporting on March 3 are going to be running versions of Firefox >= 70... except for those reporting 0, those cheeky chums.

Though most pings on that query report a null value (2,856,984 of them)?

(In reply to Dave Townsend [:mossop] (he/him) from comment #6)

Romain, do you still want this?

Sorry for the delay Dave, lots going on.
Yes I still want this, let's seefirst if there is an issue with the probe (Comment 8 shows most values are null unless I am misunderstanding the query)

The priority flag is not set for this bug.
:mossop, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(dtownsend)

(In reply to Romain Testard [:RT] from comment #8)

Apologies for the delay, a lot going on...
Regarding the request, thanks for the pointer, I attach the response.

(In reply to Chris H-C :chutten from comment #3)

I think there may be an error in that query. I'm not sure what type histogram_parent_number_of_profiles is, but it should be a STRUCT, so the GROUP BY should fail. Here's a query using telemetry.main instead of main_summary which looks at the sum of the histogram:

https://sql.telemetry.mozilla.org/queries/69110/source#174377

The values are about what I'd expect given that most users reporting on March 3 are going to be running versions of Firefox >= 70... except for those reporting 0, those cheeky chums.

Though most pings on that query report a null value (2,856,984 of them)?

Yes, that's because most pings cover timespans that don't include startup. Startup's the time that the value is recorded, so the first ping after startup will contain that value. All subsequent pings until the next startup will have no value.

Comment on attachment 9137041 [details] renewal_request.md Is the provided Data Collection Review complete, correct, and data-review+ by a Data Steward? Yes. Is the data collection covered by the existing Firefox Privacy Notice? Yes. --- Result: datareview+
Attachment #9137041 - Flags: data-review?(chutten) → data-review+

Taking to update schema in tree.

Assignee: nobody → dtownsend
Flags: needinfo?(dtownsend)

The priority flag is not set for this bug.
:mossop, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(dtownsend)
Pushed by dtownsend@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/c2d292ae8944 Re-enable the NUMBER_OF_PROFILES telemetry. r=Gijs
Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla76
Flags: needinfo?(dtownsend)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: