1284932 - Investigate Fennec Telemetry retention mismatch with Adjust data

Status: RESOLVED INCOMPLETE

(Cloud Services Graveyard :: Metrics: Pipeline, defect, P3)

Description

[Georg Fritzsche [:gfritzsche]]

This is following up to the findings from bug 1260715.
We need to understand why the retention numbers between Adjust & Telemetry is off by a big amount.

* unsorted London notes:
  https://docs.google.com/document/d/1cYGaQ3s2Vhk489oi-bLPtwtmLkNZqsRA_QhzWk6bvCY/
* re-structured docs of the issues & status:
  https://docs.google.com/document/d/17sxCUFo43F-z8wx2FVoDsKBBnHBwpLTcaM330tL1790/

Comment 1

[Thomas Huelbert]

to you Georg, please re-assign/unassign if that wrong.

Comment 2

[Georg Fritzsche [:gfritzsche]]

It was scheduled for this iteration, i'll assign myself to it when i'm back on this.

Comment 3

[Georg Fritzsche [:gfritzsche]]

In a first step, i've updated the new document with charts for active users and retention:
https://docs.google.com/document/d/17sxCUFo43F-z8wx2FVoDsKBBnHBwpLTcaM330tL1790/

Active user numbers seem fine from a high-level look. They are trending towards 10% below the Adjust numbers, which seems reasonable given update lag (we only get proper UT data from Fennec 46 on).

Retention is definitely off:
* d0 (or new_records) is way higher in UT than in Adjust (trending towards UT being 140% of Adjust)
* d1 is less of, trending towards UT being 124% of Adjust
* d7 and d30 are less off respectively

Comment 4

[Thomas Huelbert]

please update once you start working on it FRank

Comment 5

[Frank Bertsch [:frank]]

I'll be working on this, if needed, after the KPI dashboard reboot.

Comment 6

[Firefox Bug Husbandry Bot]

Closing abandoned bugs in this product per https://bugzilla.mozilla.org/show_bug.cgi?id=1337972