1282587 - Remove SCROLL_LINKED_EFFECT_FOUND telemetry probe (or extend the expiry)

Status: RESOLVED FIXED

(Core :: Panning and Zooming, defect, P3)

Description

[Kartikaya Gupta (email:[e-mail])]

Got this email:

===
The following histograms will be expiring on 2016-08-01, and should be removed from the codebase, or have their expiry versions updated:

* SCROLL_LINKED_EFFECT_FOUND expires in version 50.0a1 (watched by [e-mail]) - Attempt to determine prevalence of scroll-linked effects on the web.

This is an automated message sent by Cerberus. See https://github.com/mozilla/cerberus for details and source code.
===

I think we can probably remove this probe now, the number it's been reporting have been pretty stable and we don't need to keep it in the tree.

Comment 1

[Kartikaya Gupta (email:[e-mail])]

Attachment: Bug 1282587 - Remove telemetry probe expiring in Firefox 50.

Review commit: https://reviewboard.mozilla.org/r/60910/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/60910/

Comment 2

[Botond Ballo [:botond]]

Comment on attachment 8765623 [details]
Bug 1282587 - Remove telemetry probe expiring in Firefox 50.

https://reviewboard.mozilla.org/r/60910/#review58178

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/1dc9bc2129e8
Remove telemetry probe expiring in Firefox 50. r=botond

Comment 4

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/1dc9bc2129e8

Comment 5

[Karl Dubost💡 :karlcow]

In https://bugzilla.mozilla.org/show_bug.cgi?id=1331312#c34

kats said:
> We had such telemetry, added in bug 1229052, and then subsequently removed in
> bug 1282587 because the number it had been reporting was pretty stable.

What were the numbers like?
And what was measured? :)
(sorry if I'm asking a dumb question, just want to better understand.)

Comment 6

[Kartikaya Gupta (email:[e-mail])]

It looks like telemetry.m.o doesn't show data for old probes. But the value was in the neighbourhood of 4% on desktop and 5% on mobile, if I recall correctly. I dug out the old email thread referenced at [1] and in that I claimed that  SCROLL_LINKED_EFFECT_FOUND was true for 4.2% of pages on beta 46.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=1266874#c1

Comment 7

[Kartikaya Gupta (email:[e-mail])]

Oh, and as to "what was measured" - for each top-level document the browser loaded, we recorded whether or not there was a change to certain CSS properties made by scroll event listeners. The CSS properties are the ones at [1]. This was used as an indication of whether or not the page was running JS to try and keep things visibly in-sync with the scroll position. It's not a perfect heuristic but I think good enough to be useful and representative.

[1] http://searchfox.org/mozilla-central/rev/31b6089ce26fa76459642765115605d50a6c67b4/layout/style/nsDOMCSSDeclaration.cpp#63-79