Add the stats tracked in T246104: Track opt-out user preference for new conflict UI to Grafana.
https://grafana.wikimedia.org/d/000000346/mediawiki-twocolconflict
Add the stats tracked in T246104: Track opt-out user preference for new conflict UI to Grafana.
https://grafana.wikimedia.org/d/000000346/mediawiki-twocolconflict
Data comes in waiting for demo, see https://grafana.wikimedia.org/d/000000346/mediawiki-twocolconflict?orgId=1&refresh=1d&from=now-7d&to=now&fullscreen&panelId=48
To do: check wether this is showing duplicates of users opting out in different wikis.
Yes, the current implementation of this is showing duplicates of users when they opt-out on different wikis. I'm not sure how easy this can be resolved.
Change 587232 had a related patch set uploaded (by WMDE-Fisch; owner: WMDE-Fisch):
[analytics/wmde/scripts@master] Only track unique users disabling TwoColConflict
Change 587232 merged by jenkins-bot:
[analytics/wmde/scripts@master] Only track unique users disabling TwoColConflict
Change 587496 had a related patch set uploaded (by WMDE-Fisch; owner: WMDE-Fisch):
[analytics/wmde/scripts@production] Only track unique users disabling TwoColConflict
Just for the record: We decided on just counting the unique users with the setting disabled.
Change 587496 merged by jenkins-bot:
[analytics/wmde/scripts@production] Only track unique users disabling TwoColConflict
Starting tomorrow the numbers should only reflect unique users that have the setting disabled.
Change 587720 had a related patch set uploaded (by WMDE-Fisch; owner: WMDE-Fisch):
[analytics/wmde/scripts@master] Fix unique user count for TwoColConf
Change 587720 merged by jenkins-bot:
[analytics/wmde/scripts@master] Fix unique user count for TwoColConf
Change 587728 had a related patch set uploaded (by WMDE-Fisch; owner: WMDE-Fisch):
[analytics/wmde/scripts@production] Fix unique user count for TwoColConf
Change 587728 merged by jenkins-bot:
[analytics/wmde/scripts@production] Fix unique user count for TwoColConf
Opt-out counts are correctly tracking the unique users now. See the corresponding pane form the board[1].
The gap on the 9th was due to a bug in the first version of that patch. But now it's a good hint when the unique counts started. On both sides of the gap the number is the same so I guess it's likely that the numbers before were unique as well by accident.
[1] https://grafana.wikimedia.org/d/000000346/mediawiki-twocolconflict?panelId=48&fullscreen&orgId=1