Closed Bug 1603691 Opened 4 years ago Closed 4 years ago

Extra telemetry events are generated after few minutes for Doh 1.3.0

Categories

(Firefox :: Security, defect, P2)

72 Branch
defect

Tracking

()

RESOLVED DUPLICATE of bug 1654520
Tracking Status
firefox72 - wontfix
firefox73 --- affected

People

(Reporter: mberlinger, Assigned: nhnt11)

References

Details

[Affected Platforms]:

  • Windows 10x64
  • Windows 7x64
  • macOS 10.15
  • Ubuntu 18.04x64

[Affected Versions]:

  • Firefox Beta 72.0b5

[Steps to reproduce]:

  1. Open the browser.
  2. Go to about:config and add doh-rollout.enabled to true.
  3. Go to about:telemetry#events-tab.
  4. Wait ~1min.

[Expected result]:

  • No extra telemetry events generated

[Actual result]:

  • After ~1min another telemetry event is received

[Regression range]:

  • This issue isn’t is a regression since the 1.3.0 Doh is implemented as in-tree add-on just from Firefox 72 Beta

[Notes]:

  • If disconnecting and reconnecting to network more telemetry events are generated with netChange status and one with startup on Windows
  • If disconnecting and reconnecting to network more telemetry events are generated with startup status and one with netChange on Ubuntu
  • On macOS no netChange telemetry is generated only startup
  • This is happening also when having Safe Search set
  • The issue is reproducible with the 1.3.0 in-tree add-on.

Nihanth, do you know about this?

Component: General → Security
Flags: needinfo?(nhnt11)

Sounds like this may be a blocker for the feature or at least, P1.

I'm aware of this one, we fixed it in Github already and need to port the change into the m-c version.

Assignee: nobody → nhnt11
Status: NEW → ASSIGNED
Flags: needinfo?(nhnt11)
Priority: -- → P1

Is there an ETA for porting the change over? Is this fix a must-have for 72?

Flags: needinfo?(nhnt11)

Wayne do you have an answer for comment 4?

Flags: needinfo?(wthayer)

Hey, I'm going to make a patch for this in the next few hours. Sorry for the delayed response. Ideally, this should be uplifted to 72.

Flags: needinfo?(nhnt11)
Flags: needinfo?(wthayer)

Hello,

Ideally we would need a Fx 72 beta build (treeherder build) by tomorrow morning 8 AM (EET). Due to the limited bandwidth we can only accommodate spotchecks for this fix as it was a pretty big blocker for us and running all the blocked tests is not doable. If this is not something that we (QA + Engineering) are comfortable with we can't sign-off this feature as Green/Yellow.

Feel free to reach out on slack when the build is finished or if anything pops-up.

Flags: needinfo?(nhnt11)

I don't think this is a blocker. We run heuristics one more time than necessary, but the extra result doesn't have any harmful consequences. It is a bug and should be fixed, but not a blocker for 72.

I dug into this for a while, and it turns out that we need to fix captive portal handling (bug 1600109) to fix this "properly".

I'm discarding my attempt at this and switching over to that bug instead. Once that's done, this bug should be solved, but I'll likely use this bug to land some telemetry tests before resolving.

Flags: needinfo?(nhnt11)
Depends on: 1600109
Priority: P1 → P2

Daniel, just to be very clear, this will not need testing in the morning.

Cleaning some stuff up. Marking this a dupe of bug 1654520 which likely fixed or greatly mitigated the situation and if there's still an issue we can re-open or file a new bug.

Status: ASSIGNED → RESOLVED
Closed: 4 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.