1200944 - Instrument UI Pref for TP in PBM with Telemetry probe

Status: RESOLVED FIXED

(Firefox :: General, defect, P1)

Description

[Javaun Moradi [:javaun]]

There is a UI editable pref under Preferences > Privacy that enables TP in PBM. It is on by default in DevEdition now.

This is the pref:
privacy.trackingprotection.pbmode.enabled

We'd like to instrument this for telemetry so we can understand how many people globally disable TP. This may tell us: 1. how many people dislike TP in PBM, or 2. that our on/off switch is really hard to find. There is also now an on/off switch on the PBM built-in page itself. We just want to instrument the pref and can't know where it's been changed. We can't instrument anything that would reveal a user went to PBM

Comment 1

[Panos Astithas (he/him) [:past] (please ni?)]

Attachment: Patch v1

I've reversed the logic in this probe because TP in PBM is enabled by default, whereas normal TP is disabled by default.

Comment 2

[Allison Naaktgeboren :ally]

Comment on attachment 8661734 [details] [diff] [review]
Patch v1

Review of attachment 8661734 [details] [diff] [review]:
-----------------------------------------------------------------

So with the exception of the expiration version, the probe looks fine.

The bigger issue is that we usually do not allow the collection about user behavior in private browsing mode even if the user has opted into telemetry. I understand why we want to know about when our new feature is disabled and I want us to be consistent in firefox about what we track and when.

Unfortunately, I'll need to talk to the other data stewards and get back to you about it.

::: toolkit/components/telemetry/Histograms.json
@@ +8072,5 @@
>      "kind": "boolean",
>      "description": "Whether or not a session has tracking protection enabled"
>    },
> +  "TRACKING_PROTECTION_PBM_DISABLED": {
> +    "expires_in_version": "never",

We do not allow "never" as a value for expiration. Probes can always be renewed if they are still in use by their expiry version. TP is a new feature, so I would version 60 should be plenty of time for it to be great..or y'know.

Comment 3

[Allison Naaktgeboren :ally]

Comment on attachment 8661734 [details] [diff] [review]
Patch v1

NI myself to followup up with other stewards

Comment 4

[Panos Astithas (he/him) [:past] (please ni?)]

(In reply to Allison Naaktgeboren please NEEDINFO? :ally from comment #2)
> The bigger issue is that we usually do not allow the collection about user
> behavior in private browsing mode even if the user has opted into telemetry.
> I understand why we want to know about when our new feature is disabled and
> I want us to be consistent in firefox about what we track and when.

But isn't that what the PrivateBrowsingUtils.isWindowPrivate(window) check is for? The probe is sent only when the pref modification is detected in a normal window, which, to be honest, is where I would expect most users to flip the pref.

Comment 5

[Marco Mucci [:MarcoM]]

Hi Panos, should this bug be marked for QE verification?

Comment 6

[Panos Astithas (he/him) [:past] (please ni?)]

No, I think we don't QA telemetry changes.

Comment 7

[Allison Naaktgeboren :ally]

(In reply to Panos Astithas [:past] from comment #4)
> (In reply to Allison Naaktgeboren please NEEDINFO? :ally from comment #2)
> > The bigger issue is that we usually do not allow the collection about user
> > behavior in private browsing mode even if the user has opted into telemetry.
> > I understand why we want to know about when our new feature is disabled and
> > I want us to be consistent in firefox about what we track and when.
> 
> But isn't that what the PrivateBrowsingUtils.isWindowPrivate(window) check
> is for? The probe is sent only when the pref modification is detected in a
> normal window, which, to be honest, is where I would expect most users to
> flip the pref.

Oh, then you're fine. If you do want to track turning it off in private browsing, the stewards are ok with it but only for a version or two.

Comment 8

[Panos Astithas (he/him) [:past] (please ni?)]

That's fine, comment 0 clearly indicates that we don't really want to know specifically when TP is turned off in PBM. I'll land this with the expiration change requested in comment 2.

Comment 9

[Pulsebot]

https://hg.mozilla.org/integration/fx-team/rev/58adc6b9c8ea

Comment 10

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/58adc6b9c8ea