Telemetry Alert for CHECKERBOARD_DURATION on 2019-05-11
Categories
(Core :: Panning and Zooming, defect, P3)
Tracking
()
People
(Reporter: kats, Unassigned)
Details
From the email:
We have detected a change in the Telemetry probe CHECKERBOARD_DURATION in Firefox Nightly builds from 2019-05-11.
Alert details: http://alerts.telemetry.mozilla.org/index.html#/detectors/1/metrics/1717/alerts/?from=2019-05-11&to=2019-05-11
Changes new to Nightly builds on 2019-05-11: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=2abcefb31ba7b2a1c573edc5695b772826c6a078&tochange=4440662cd0177f3fc3200cff190daed47d71a167
The value of CHECKERBOARD_DURATION over time: https://telemetry.mozilla.org/new-pipeline/evo.html#!measure=CHECKERBOARD_DURATION
|
Reporter | |
Comment 1•6 years ago
|
||
So the change seems to be improvement (graph shifts left, towards smaller duration numbers). So yay!
I have no idea what caused this; nothing in the pushlog range really stands out as being able to cause this sort of a dramatic effect.
|
|
Reporter | |
Comment 2•6 years ago
|
||
Hm actually it looks like the sample count went up at the same time. So instead of being an improvement, what happened was something introduced a new short-duration checkerboard that probably happens on startup and so it brought down the median.
|
|
Reporter | |
Comment 3•6 years ago
|
||
I did a local windows build that logged checkerboard instances and ran it, hoping that it would hit the startup checkerboard and then I could figure out where it was coming from, but no dice. My local build doesn't seem to hit it.
At any rate, while this is technically a regression I suspect it's not going to impact the user-observed behaviour very much.
|
|
Reporter | |
Comment 4•5 years ago
|
||
Not actionable
Description
•