[10:06:14] <dcausse>	 lunch
[13:07:52] <pfischer>	 dcausse: I have a question regarding your review comment (https://gitlab.wikimedia.org/repos/search-platform/cirrus-streaming-updater/-/merge_requests/5#note_42832): Would that break or silently ignore all updates of a bulk request? In my integration test I mix super_detect_noop and painless updates and get a 200 w/o any errors.
[13:09:27] <dcausse>	 pfischer: my bad I was not very precise, I meant merging two events into one, not combining multiple updates in a bulk request
[13:11:39] <dcausse>	 say we wanted (I don't think we have a compeling use-case yet) to combine for the same page: a rev_update + redirect update this could be possible with the noop_handler
[13:25:54] <pfischer>	 Alright, I’ll look into adding a max size to the set handler then.
[15:02:20] <dcausse>	 changed the "cirrus failures" graph in https://grafana-rw.wikimedia.org/d/000000455/elasticsearch-percentiles?orgId=1&from=now-7d&to=now&refresh=1m, the ops/minute on graphite .count was wrong
[15:04:31] <dcausse>	 we reject 2.5 fulltext queries per sec because someone is sending a broken query in a loop.... "&& John Doe, Jean Dupont, random names" (always the same query...)
[15:44:02] <dcausse>	 going offline early, back later tonight