Closed Bug 1240887 Opened 10 years ago Closed 8 years ago

Compare BHR rates vs chrome-hang rates on Nightly

Categories

(Toolkit :: Telemetry, defect)

Product:
Toolkit
Component:
Telemetry
Version:
46 Branch
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Tracking Flags:
Tracking Status
firefox46 --- affected

People

(Reporter: vladan, Unassigned)

References

Details

User Story

We need to increase our confidence in the correctness of the BHR measurement. :billm had the idea of comparing BHR hang rates vs chrome-hang hang rates.

There are some gotchas with this analysis:

1. Only Nightly on Windows has chrome-hang reporting, so we'll have to limit ourselves to the data collected from Nightly 46 on Windows
2. Bug 1234618 badly messed up BHR reliability, so we'll only want to look at Nightly buildIDs >= 20151224 which have the fix
3. Chrome-hangs are only reported if the hang lasts longer than 5 seconds, so we should only count BHR hangs ythat are 4095ms or longer.

So ultimately, we'll be comparing BHR hang rates (for hangs >=4 seconds) vs chrome-hang rates only for Nightly 46 on Windows for buildIDs > 20151224.

This should give us a rough idea of how reliable BHR reporting is (for hangs > 4 seconds).
Roberto: can you & poiru do this analysis?
Flags: needinfo?(rvitillo)
Flags: needinfo?(birunthan)
User Story: (updated)
Assignee: nobody → birunthan
Status: NEW → ASSIGNED
Flags: needinfo?(rvitillo)
Flags: needinfo?(birunthan)
Poiru's analysis of Nightly chromehangs & BHR hangs (not normalized by # of sessions yet): https://gist.github.com/poiru/961e767a08a07161c175
At first glance, it looks like BHR & chromehang counts are far apart
Assignee: birunthan → nobody
Status: ASSIGNED → NEW
Hangs were very closely scrutinized during the e10s rollouts.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.