Find a way to make /daily work with new "Android" reporting in the short term

VERIFIED FIXED in 22

Status

task
--
critical
VERIFIED FIXED
7 years ago
7 years ago

People

(Reporter: kairo, Assigned: selenamarie)

Tracking

unspecified
ARM
Android
Dependency tree / graph

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [qa-])

Reporter

Description

7 years ago
If you look at  https://crash-stats.mozilla.com/daily?form_selection=by_version&p=FennecAndroid&v[]=18.0a1&v[]=17.0a2&v[]=16.0b4&v[]=15.0.1&hang_type=any&os[]=Windows&os[]=Mac+OS+X&os[]=Linux&date_range_type=build&date_start=2012-09-13&date_end=2012-09-27&submit=Generate you'll notice that starting with 2012-09-26 this reports you no users.

This is because the new Breakpad version we have in there now reports the OS to be "Android" instead of Linux, see a sample crash report at bp-d0213091-5c3d-42ef-a747-04f942120927 and so they don't get counted against "Linux" there any more. Note that ADUs still get reported as "Linux" though,so this doesn't match.

We should do a short-term fix here to have them somehow show up as "Linux" in the UI (I guess the fix might actually be MWare or DB but I'll leave that up to you guys), until we have the possibility to show Android as Android in future UI.
Selena,

Adding a row to os_name_matches would map Android to Linux.
Assignee: nobody → sdeckelmann
Reporter

Comment 2

7 years ago
Can we please get some progress here?
This is significantly impairing our ability to assess Android stability, and the change that caused this will also make its way into Aurora early next week.
Target Milestone: --- → 22
Severity: normal → critical
Did this on .stage: 

 insert into os_name_matches values('Linux', 'Android%'); 

Will try to verify that this had the intended effect on /daily per KaiRo's IRC comments.
Depends on: 797566
(In reply to Selena Deckelmann :selena from comment #3)
> Did this on .stage: 
Just checking in -- is the bug ready to be bumped to fixed?
Reporter

Comment 5

7 years ago
(In reply to Selena Deckelmann :selena from comment #3)
> Did this on .stage: 
> 
>  insert into os_name_matches values('Linux', 'Android%'); 
> 
> Will try to verify that this had the intended effect on /daily per KaiRo's
> IRC comments.

Seems not to be the case. :(

This is especially visible on https://crash-stats.allizom.org/daily?form_selection=by_version&p=FennecAndroid&v[]=18.0a1&v[]=17.0a2&v[]=16.0b6&v[]=15.0.1&hang_type=any&os[]=Windows&os[]=Mac+OS+X&os[]=Linux&date_range_type=build&date_start=2012-09-23&date_end=2012-10-07&submit=Generate - we should have numbers for crashes on 18.0a1 for all days, but due to this bug, it goes to 0 starting with 2012-09-26.
This may just be because stage got rebuilt without the data change. 

I'm re-adding it and re-running backfill_matviews to see if this works now.
https://github.com/mozilla/socorro/pull/872 -- solves problems going forward, but to fix old reports, we need to do some backfilling
Status: NEW → ASSIGNED
Whiteboard: [qa-]
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
To prove this worked with a recent matview run: 

select uuid, os_name from reports_clean where uuid in ( select uuid from reports where date_processed > '2012-10-10' and os_name = 'Android' limit 1);

os_name should be "Linux"
Reporter

Comment 10

7 years ago
Verified by the data for yesterday having the Android data in the Linux columns.
Status: RESOLVED → VERIFIED
Great success -- thank you kairo.
You need to log in before you can comment on or make changes to this bug.