Page MenuHomePhabricator

Rerun Movement Metrics Puppet Job after Issues in Mediawiki_history dumps are fixed
Closed, InvalidPublic

Description

Because of the ongoing pause in production of mediawiki_history snapshots (T377594) available in wmf.mediawiki_history we will need to rerun the movement metric job that create our editor datasets in wmf_product for the October 2024 snapshots.

This should be done after T377594 is completed, dumps data is QA'd, and data is backfilled in wmf.mediawiki_history

Event Timeline

OSefu-WMF triaged this task as Medium priority.

Actually, unless I'm missing something, we don't need to do anything here.

The dumps and the downstream mediawiki_wikitext_history are paused, but mediawiki_history is still fine (obviously, there's nothing at all confusing about those names 😆). Our movement metrics intermediate tables only build from mediawiki_history, so everything's fine there. Our dependency on mediawiki_wikitext_history is only through Fabian's content gaps tables, so he's in charge of rebuilding those once the dumps are running again.

Actually, unless I'm missing something, we don't need to do anything here.

The dumps and the downstream mediawiki_wikitext_history are paused, but mediawiki_history is still fine (obviously, there's nothing at all confusing about those names 😆). Our movement metrics intermediate tables only build from mediawiki_history, so everything's fine there. Our dependency on mediawiki_wikitext_history is only through Fabian's content gaps tables, so he's in charge of rebuilding those once the dumps are running again.

You're 100% right. My mistake. I see the Oct snapshots in wmf.mediawiki_history