Nothing Special   »   [go: up one dir, main page]

Page MenuHomePhabricator

Standardize definition of mentee that show in the mentor-facing features
Open, Needs TriagePublic

Description

Currently, the mentee overview table in GrowthExperiments-MentorDashboard shows different users than the filters in recent changes. This is confusing, and resulted in a couple of tasks filled (see subtasks).

We should:

  1. Decide which kind of mentees we want to show to mentors (currently, mentee overview shows users who made an edit in last 6 months or registered in last 14 days; we might want to add an experience limit, and we likely should account for opt-outs from T287915)
  2. Ensure both recent changes and mentee overview show the same set of users
  3. Ensure recent changes filters stay realtime

During the 2022-10-26 mentorship meeting, we decided to include users who registered in the last 6 months in both the Mentor dashboard and recent changes filters. From the requirements above:

  1. We'll show users who registered in last 6 months (ie. display more users in Mentee overview)
  2. All users who show in recent changes will show in Mentee overview (once it updates); all users who appear in Mentee overview will show in Recent changes (when they make an action, of course)
  3. We will not change the implementation of the RC filters, so that should stay as-is. T318457 will bring it to the biggest wikis, where it is not yet available.

Related Objects

Event Timeline

Urbanecm_WMF updated the task description. (Show Details)

In the 2022-10-26 mentorship meeting, we decided to prioritize this. I'll look into how much this will inflate growthexperiments_mentee_data. If the growth is acceptable, the change will be done.

Data for a couple of our wikis are below (source):

image.png (255×610 px, 19 KB)

Looks like we're looking at ~250% of rows in the growthexperiments_mentee_data table (the number of rows should stabilize post-change, as rows are purged regularly). Is that acceptable?

Complete numbers for all affected wikis (average growth is to 330%):

wikishould_be_in_dashboardis_in_dashboarddiffdiff_%
ar.wikipedia.org557531795937794310.45
ary.wikipedia.org541125.0
az.wikipedia.org23509191431255.71
bn.wikipedia.org1120727118496413.39
ckb.wikipedia.org753283470266.08
cs.wikipedia.org792932834646241.52
de.wikipedia.org393501513224218260.04
el.wikipedia.org359513422253267.88
en.wikipedia.org8226821276547637386.67
es.wikipedia.org30976730723669423.92
eu.wikipedia.org862617245139.71
fa.wikipedia.org297171198217735248.01
frr.wikipedia.org10191000.0
fr.wikipedia.org689603143337527219.39
fr.wiktionary.org16423751267437.87
he.wikipedia.org895645384418197.36
hi.wikipedia.org110100.0
hr.wikipedia.org1219632587192.88
hu.wikipedia.org379021621628175.3
hy.wikipedia.org995476519209.03
id.wikipedia.org22630779114839290.46
it.wikipedia.org234551072112734218.78
jv.wikipedia.org361719211.76
kn.wikipedia.org18066114272.73
ko.wikipedia.org841930905329272.46
ks.wikipedia.org19613316.67
ku.wikipedia.org973859255.26
lv.wikipedia.org466232234200.86
ne.wikipedia.org1025052204.0
nl.wikipedia.org926442674997217.11
no.wikipedia.org329417541540187.8
pl.wikipedia.org1211461425972197.23
pt.wikipedia.org481801767730503272.56
ro.wikipedia.org415316622491249.88
ru.wikipedia.org540991936034739279.44
sk.wikipedia.org1598761837209.99
sq.wikipedia.org952363589262.26
sr.wikipedia.org1930934996206.64
sv.wikipedia.org544826092839208.82
test.wikipedia.org36964305576.56
te.wikipedia.org1196298898401.34
tr.wikipedia.org26342660219740399.0
uk.wikipedia.org961645715045210.37
vi.wikipedia.org1244843338115287.28
zh.wikipedia.org315971248319114253.12

Next step is to decide whether this growth is acceptable. As I noted in my previous column, the new number of rows will stabilize after a while, as old rows are purged.

Based on a discussion with @Tgr, and IRC conversation with DBA (IRC logs, see 2022-11-21 17:54:37 onwards), the growth should be fine from technical point of view.