1284945 - Extend KEYGEN_GENERATED_KEY_TYPE until Firefox 55

Status: RESOLVED FIXED

(Core :: Security: PSM, defect, P1)

Description

[Kate McKinley [:kmckinley, :Kate]]

We need continue to collect data to determine whether we should remove keygen, keep as-is, or update it in some way.

Comment 1

[Kate McKinley [:kmckinley, :Kate]]

Attachment: Bump the version number

Allison, do we need a privacy review if all we are doing is bumping the version number?

Comment 2

[Dana Keeler (she/her) (use needinfo) [:keeler]]

Comment on attachment 8768474 [details] [diff] [review]
Bump the version number

Review of attachment 8768474 [details] [diff] [review]:
-----------------------------------------------------------------

FWIW, r=me. Not sure if a telemetry peer needs to sign off on this. I assume Ally will tell us.

Comment 3

[Kate McKinley [:kmckinley, :Kate]]

Attachment: Bump the version number

Comment 4

[Allison Naaktgeboren :ally]

Comment on attachment 8768534 [details] [diff] [review]
Bump the version number

Review of attachment 8768534 [details] [diff] [review]:
-----------------------------------------------------------------

Apologies for the delay, I have been sick.

Yes, you do need a review from data collection to extend collection, especially release collection ones.

From a technical perspective, please add the bug number field to the probe since you are updating it. Bonus points if you also add Bug 1191414, to that list, as that's the original bug for this probe as far as I can tell.

Release collection probes have a higher bar of justification. Please explain why the previous data collection for from the last 10 months or so hasn't been sufficient and why 10 more versions will. 

I'm not adverse to allowing a bump, but it needs to be justified. I promise I'm not trying to make your lives any more difficult than need be.

Comment 5

[Allison Naaktgeboren :ally]

See review comment above.

Comment 6

[Dana Keeler (she/her) (use needinfo) [:keeler]]

One of the concerns we've had around removing keygen is that usage might not be uniformly distributed in time. That is, there might be parts of the year that have a huge spike in usage (e.g. for voting or submitting taxes). Having at least full year's worth of data will go a long way towards either confirming or disproving this theory, and in any case it will give us a better idea of whether or not we can remove keygen without negatively impacting too many users.

Comment 7

[Allison Naaktgeboren :ally]

(In reply to David Keeler [:keeler] (use needinfo?) from comment #6)
> One of the concerns we've had around removing keygen is that usage might not
> be uniformly distributed in time. That is, there might be parts of the year
> that have a huge spike in usage (e.g. for voting or submitting taxes).
> Having at least full year's worth of data will go a long way towards either
> confirming or disproving this theory, and in any case it will give us a
> better idea of whether or not we can remove keygen without negatively
> impacting too many users.

Then let us say Firefox 55 as the expiration? That in addition to the data already collected, should give you a full year and then some. 

It also may be in your best interest to ping whoever is in charge of telemetry servers to verify that the data from this probe will be kept for a full year.  

Using expiration 55 + changes in comment 4 p=ally.

Comment 8

[Kate McKinley [:kmckinley, :Kate]]

Attachment: Bump the version number to FF55 to give a year of data

Updated expiration to version 55 and added the bug numbers.

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/e543221ad4ee
update KEYGEN_GENERATED_KEY_TYPE to collect until Firefox 60. r=keeler,p=ally

Comment 10

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/e543221ad4ee

Comment 11

[:Cykesiopka]

(Updating summary to reflect that the extension ended up being up to 55 instead of 60.)