1510359 - Document tracking OOM crashes

Status: NEW

(Data Platform and Tools :: Documentation and Knowledge Repo (RTMO), enhancement, P2)

Description

[William Lachance (:wlach)]

Currently OOM crashes are not treated any differently from others inside error_aggregates (they are just considered regular old "main" or "content" crashes) but for some use cases like mission control it would be useful to track them seperately, so we can easily see whether they are increasing or decreasing normalized over the amount of usage per release.

I believe it should be possible to auto-detect OOM keywords by looking at the contents of the crash pings, so hopefully this won't be too difficult.

Comment 1

[William Lachance (:wlach)]

I filed an issue to mission control for implementing a view of this data once it's there:

https://github.com/mozilla/missioncontrol/issues/340

Comment 2

[Andrew Overholt [:overholt]]

We often consider OOM crashes ("small" or "large") distinctly when looking at crash reports. We also sometimes want to specifically look at OOM rates (consider the default number of content processes, for example). I'm sure there are other cases where it would be nice to have them be separate from other crash types.

Comment 3

[David Lawrence [:dkl]]

Moving to General.

Comment 4

[William Lachance (:wlach)]

Probably still want some way of doing this. I think it is possible to determine this from the crash ping metadata now in telemetry.crash, maybe I can just document that somewhere like this: https://docs.telemetry.mozilla.org/cookbooks/crash_pings.html

Comment 5

[William Lachance (:wlach)]

Needinfo'ing myself to remind me to get this done. I think it would be good to be able to resolve this bug.

Comment 6

[Mark Reid [:mreid]]

+1 for adding this to the crash ping documentation on docs.tmo.