1247327 - Add WebGL acceptance rates to telemetry

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Comment 1

[Jeff Muizelaar [:jrmuizel]]

Attachment: Track whether webgl context creation succeeds

Comment 2

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/1fb660fde013

Comment 3

[Jeff Muizelaar [:jrmuizel]]

Vladan reviewed this in person.

- What are the questions that these probes are designed to address?
How many people are not getting WebGL
- Is this the minimum amount of data necessary to answer those questions?
Yes
- Who is responsible for monitoring the data?
The GFX team/Me
- How often will they do it?
Maybe once a quarter
- Do the necessary monitoring dashboards already exist or who is going to create them?
Not yet, but it should be easy to add to dvander's telemetry stuff
- What's the user value?
Hopefully more people will WebGL

Comment 4

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/1fb660fde013

Comment 5

[Milan Sreckovic [:milan] (needinfo for best results)]

Feels like there is overlap between CANVAS_WEBGL_USED and CANVAS_WEBGL_SUCCESS.  I'm not sure we need CANVAS_WEBGL_USED anymore, since the existence of either value in CANVAS_WEBGL_SUCCESS would suggest we were trying to use it.
Also, does it make sense to combine the WebGL1 and WebGL2 and just have an integer histogram that captures all of these?  WebGL1-Failed, WebGL1-Succeeded, WebGL2-Failed, WebGL2-Succeeded - a single histogram with four values instead of three histograms?

Comment 6

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 8718001 [details] [diff] [review]
Track whether webgl context creation succeeds

Review of attachment 8718001 [details] [diff] [review]:
-----------------------------------------------------------------

I'd like to get this on aurora/beta so that we can gather data sooner without waiting.

Comment 7

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 8718001 [details] [diff] [review]
Track whether webgl context creation succeeds

Approval Request Comment
[Feature/regressing bug #]: None
[User impact if declined]: None
[Describe test coverage new/current, TreeHerder]: None
[Risks and why]:  Just adds telemetry

Comment 8

[Jeff Muizelaar [:jrmuizel]]

I'll make another patch that removes CANVAS_WEBGL_USED

Comment 9

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8718001 [details] [diff] [review]
Track whether webgl context creation succeeds

Too late for 45.

Comment 10

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 8718001 [details] [diff] [review]
Track whether webgl context creation succeeds

Review of attachment 8718001 [details] [diff] [review]:
-----------------------------------------------------------------

I looked at the telemetry results for this patch and it turns out the patch is garbage.

Comment 11

[Jeff Muizelaar [:jrmuizel]]

Attachment: Try for more success with counting this

Comment 12

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/863c6740f637

Comment 13

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/863c6740f637

Comment 14

[Jukka Jylänki]

Looks great! Jeff, once we've got a bit more data on this, could you post them my way? Or if the data is public, can I visit a page to track it manually?

Comment 15

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 8724956 [details] [diff] [review]
Try for more success with counting this

Approval Request Comment
[Feature/regressing bug #]: None
[User impact if declined]: We'll have less visibility into WebGL success rates.
[Describe test coverage new/current, TreeHerder]: The data seems to be showing up correctly on telemetry.mozilla.com
[Risks and why]: Low risk as it's just a telemetry addition.

Comment 16

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Comment on attachment 8724956 [details] [diff] [review]
Try for more success with counting this

This is already in Aurora47, I believe the request was to uplift to 46 (Beta). Flipping the uplift flag to reflect that.

Comment 17

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 8724956 [details] [diff] [review]
Try for more success with counting this

Since Jeff says this now gives the correct data (in 47/48) let's uplift this for beta 4.

Comment 18

[Carsten Book [:Tomcat]]

 https://hg.mozilla.org/releases/mozilla-beta/rev/39ff5bafc618

Comment 19

[Carsten Book [:Tomcat]]

(In reply to Liz Henry (:lizzard) (needinfo? me) from comment #17)
> Comment on attachment 8724956 [details] [diff] [review]
> Try for more success with counting this
> 
> Since Jeff says this now gives the correct data (in 47/48) let's uplift this
> for beta 4.

backed out for bustage like https://treeherder.mozilla.org/logviewer.html#?job_id=923464&repo=mozilla-beta

Comment 20

[Liz Henry (:lizzard) (relman/hg->git project)]

Jeff do you want to try this again ? I am aiming to start the beta 4 build early this afternoon. So we could still get it in if you can successfully land it. Or, wontfix for beta.

Comment 21

[Liz Henry (:lizzard) (relman/hg->git project)]

We can still land this in beta 5 which will gtb on Thursday.

Comment 22

[Carsten Book [:Tomcat]]

(In reply to Liz Henry (:lizzard) (needinfo? me) from comment #21)
> We can still land this in beta 5 which will gtb on Thursday.

this is in risk missing beta 5 :(

Comment 23

[Liz Henry (:lizzard) (relman/hg->git project)]

Milan do you want to take a look here?

Comment 24

[Jeff Muizelaar [:jrmuizel]]

Crap. I forgot to land... I'll do it now.

Comment 25

[Ryan VanderMeulen [:RyanVM]]

Jeff landed this.
https://hg.mozilla.org/releases/mozilla-beta/rev/077b2fd8626d