1245575 - Clean up plugin startup histograms and probes

Status: RESOLVED FIXED

(Toolkit :: Telemetry, defect)

Description

[(No longer employed by Mozilla) Aaron Klotz]

bsmedberg pointed out to me that PLUGIN_STARTUP_MS is still defined even though it is no longer referenced in plugin code. We should remove its histogram definition and adjust the expiration of the following replacement histograms to "never":

BLOCKED_ON_PLUGIN_MODULE_INIT_MS
BLOCKED_ON_PLUGIN_INSTANCE_INIT_MS
BLOCKED_ON_PLUGIN_STREAM_INIT_MS
BLOCKED_ON_PLUGIN_INSTANCE_DESTROY_MS

As for BLOCKED_ON_PLUGINASYNCSURROGATE_WAITFORINIT_MS, we can leave that one's expiry date since async init is on hold.

Comment 1

[(No longer employed by Mozilla) Aaron Klotz]

Also, chutten tells me that BLOCKED_ON_PLUGIN_INSTANCE_INIT_MS isn't generating any data on e10s. IIRC, instance creation blocks content instead of chrome on e10s, but it still could be useful data. We should verify that the BLOCKED_ON_PLUGIN_* probes still work in that case.

Comment 2

[(No longer employed by Mozilla) Aaron Klotz]

Attachment: Histogram cleanup

Comment 3

[(No longer employed by Mozilla) Aaron Klotz]

This patch takes care of this histograms.json bit, but not the "we don't have this in e10s" part.

Comment 4

[Georg Fritzsche [:gfritzsche]]

Comment on attachment 8726276 [details] [diff] [review]
Histogram cleanup

Review of attachment 8726276 [details] [diff] [review]:
-----------------------------------------------------------------

This looks fine to me, make sure to get data collection review [0] if you hadn't checked this with Benjamin yet.

0: https://wiki.mozilla.org/Firefox/Data_Collection

Comment 5

[Chris H-C :chutten|PTO (back March 17)]

PLUGIN_INSTANCE_INIT and PLUGIN_INSTANCE_DESTROY are working in e10s and giving correct numbers (or what appears to be correct numbers). They are reported on the child process, as expected. Sorry for the confusion, I had accidentally filtered out too many submissions in my analysis.

The other measures might or might not be reported properly, not sure.

Comment 6

[(No longer employed by Mozilla) Aaron Klotz]

(In reply to Chris H-C :chutten from comment #5)
> PLUGIN_INSTANCE_INIT and PLUGIN_INSTANCE_DESTROY are working in e10s and
> giving correct numbers (or what appears to be correct numbers). They are
> reported on the child process, as expected. Sorry for the confusion, I had
> accidentally filtered out too many submissions in my analysis.
> 
> The other measures might or might not be reported properly, not sure.

If those are working then then stream ones should be okay as well. Thanks for the update.

Comment 7

[(No longer employed by Mozilla) Aaron Klotz]

https://hg.mozilla.org/integration/mozilla-inbound/rev/31655275a522c6a2696141df8466e6ece384ea96
Bug 1245575: Remove PLUGIN_STARTUP_MS histogram and update expiry settings for BLOCKED_ON_PLUGIN* histograms; r=gfritzsche

Comment 8

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/31655275a522