Closed Bug 1601080 Opened 4 years ago Closed 4 years ago

When opening an updated build after 4AM, the "metrics" ping is sent with the previous build's metrics but the current' build's client_info.

Categories

(Data Platform and Tools :: Glean: SDK, enhancement, P3)

enhancement

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: chutten, Unassigned)

References

Details

(Whiteboard: [telemetry:glean-rs:backlog])

When trying to get clients with or without a given metric in the "metrics" ping, it is difficult to slice nicely because the first "metrics" ping after updating contains the previous build's metrics but the current build's client_info.

We should bake the client_info somehow into the scheduler so that metrics from a build are associated with that build id in its "metrics" pings.

(In reply to Chris H-C :chutten from comment #0)

When trying to get clients with or without a given metric in the "metrics" ping, it is difficult to slice nicely because the first "metrics" ping after updating contains the previous build's metrics but the current build's client_info.

We should bake the client_info somehow into the scheduler so that metrics from a build are associated with that build id in its "metrics" pings.

I think this would still make things a bit complicated: if the metrics ping gets sent after a while Fenix is running, then we end up collecting metrics from before and after the update, in the same ping.

I think we should probably do something different: what if we had a way to detect that the application was updated and cut a metrics ping at startup?

Type: defect → enhancement
Flags: needinfo?(chutten)
Priority: -- → P3
Whiteboard: [telemetry:glean-rs:m?]

That would work wonderfully well. In absence of data showing one way or another, I expect this is similar behaviour to the likely update case (app updates overnight, user opens app after 4AM) so it wouldn't really even be that much of a change.

And "having a way to detect" is closely correlated with "keeping the old *_info fields handy" we'd need to do in order to send the ping with the proper values, so that's nice : )

Flags: needinfo?(chutten)
Depends on: 1602824
Whiteboard: [telemetry:glean-rs:m?] → [telemetry:glean-rs:backlog]

I think that the PR attached to Bug 1602824 actually resolves this bug too since it adds a check for new version and sends a metrics ping if there is.

Yep, Bug 1602824 fixed this bug!

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