Enable VIDEO_PLAY_TIME probe for release channel users as an opt-out probe after it is renewed in bug 1356046 .
Categories
(Core :: Audio/Video, task, P2)
Tracking
()
Tracking | Status | |
---|---|---|
firefox70 | --- | fixed |
People
(Reporter: mconley, Assigned: mconley)
References
Details
Attachments
(2 files)
47 bytes,
text/x-phabricator-request
|
Details | Review | |
2.41 KB,
text/plain
|
bmiroglio
:
data-review+
|
Details |
Something that tdsmith, astevenson and I were talking about in Whistler was whether or not we can determine if Firefox users watch more video when the Picture-in-Picture feature is enabled.
I started digging into what it would take to add a Telemetry probe to measure video playback time, and was pleased to find out that such a probe already exists - it's just expired, and has been since Firefox 63.
This bug is about reviving that probe to help measure the impact (if any) Picture-in-Picture has on video playback time.
Assignee | ||
Comment 1•6 years ago
|
||
This only enables the probe on Firefox Desktop, and only for content processes. We're
reviving this probe mainly as a way of determining if Picture-in-Picture causes users
to watch more video.
Assignee | ||
Updated•6 years ago
|
Comment 2•6 years ago
|
||
Check bug 1356046 instead.
However, tbh, I seriously doubt this probe will achieve what you describe. What would you be comparing things with? We have no comparison point and as telemetry isn't available by default in release, we have seen in the past that the numbers are so skewed towards one category of users that it makes any analysis impossible
I agree that determining that information would be great, but reviving that telemetry key alone won't do it.
Updated•6 years ago
|
Comment 3•6 years ago
|
||
telemetry isn't available by default in release
It doesn't answer your critique, but we should also start collecting this in release if we're interested in it.
Note that probe renewal is subject to a new data-collection review.
Assignee | ||
Comment 4•6 years ago
|
||
However, tbh, I seriously doubt this probe will achieve what you describe. What would you be comparing things with? We have no comparison point and as telemetry isn't available by default in release, we have seen in the past that the numbers are so skewed towards one category of users that it makes any analysis impossible
Thanks, jya. One way we might do this is to try to break some part of either the Beta or Release populations into two groups, one with PiP enabled, and one without, to see whether or not that reveals a signal.
I'm glad to see that dminor is already working on renewing the probe in bug 1356046! Perhaps I'll repurpose this one for requesting a data-review to getting it enabled to release by default.
Comment 5•6 years ago
|
||
Yes lets wait for bug 1356046 to land.
From my point of view we want a playback length probe in Release (because otherwise the values are not reliable) - which could be either this or a new one (see below).
jya: is VIDEO_PLAY_TIME_MS good enough going forward or do want/need to update this probe (or create a new one) which records how many frames got rendered (taking into consideration the frame rate of the material)?
Updated•6 years ago
|
Assignee | ||
Comment 7•6 years ago
|
||
Comment 8•6 years ago
|
||
![]() |
||
Comment 9•6 years ago
|
||
Updated•6 years ago
|
Comment 10•6 years ago
|
||
![]() |
||
Comment 11•6 years ago
|
||
bugherder |
Description
•