Metrics on the "metrics" ping used in onboarding experiments underreport from single-session uses
Categories
(Toolkit :: Telemetry, task, P1)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox161 | --- | fixed |
People
(Reporter: chutten, Assigned: chutten)
References
(Blocks 1 open bug)
Details
(Whiteboard: [fog-migration])
Attachments
(1 external file)
In https://docs.google.com/document/d/1o0X7rjXx_z-l9bK7ZlViuqwPl46yp4p1EQx8XP9mq6g/edit?tab=t.0 :jlockhart identified 36 metrics submitted on the "metrics" ping that correspond to Legacy probes that have been used in Legacy Jetstream analyses focusing on Firefox Desktop Onboarding. Onboarding experiment analyses are particularly sensitive to single-session uses of Firefox Desktop, and the "metrics" ping's design of being "Big and Slow" is particularly unsuited to report on these uses. This mismatch is a problem for conducting Onboarding experiments with Glean-sent data.
The solution we proposed in the document is to temporarily also report these 36 metrics on the "events" ping which is far more suited to report on these uses (especially after bug 2060890 and friends accelerated early "events" ping submission). Trying to shift the entire instrumentation and analysis culture to "If you want to look at it for Onboarding, put it on the "events" ping" isn't our desired end-goal, but is a the best (effort, timing, reversibility, etc.) way to unblock FOG Migration while we look for other, better designs next year.
| Assignee | ||
Updated•22 days ago
|
| Assignee | ||
Comment 1•22 days ago
|
||
Updated•21 days ago
|
Comment 3•16 days ago
|
||
| bugherder | ||
Updated•10 days ago
|
Description
•