As a user, when there is WDQS streaming updater maintenance I want less than 10 min of lag.
As a maintainer of the wdqs streaming updater I want the StateExtractionJob to run in a reasonable amount of time so that I don't have to downtime the flink streaming updater application for too long.
As of today running the state extraction job on an existing savepoint takes 6hours, this is way too slow and something is probably going wrong. Writing a similar size savepoint from CSV files take less than one minute.
AC:
- StateExtractionJob runs in less than 10minutes for wikidata