[18:03:44] <paladox>	 duesen: we managed to resolve the parsoid warmup issue it looks like
[18:04:49] <paladox>	 we were using a script within the prometheus redis exporter to gather metrics for the job queue which caused things to halt (and especially more prevalent when there's a sudden increase in job queue)
[18:05:28] <paladox>	 I've reworked things that improved the performance quite considerably (by using prometheus-node exporter node.d directory). Also managed to fix memory and cpu usage.
[18:05:50] <paladox>	 But the one thing is that the root jobs just build up and up (never deleted)
[18:06:09] <paladox>	 the comment at the top indicates it's deleted when the job is ran
[22:50:55] <Vulpix>	 I'm a bit surprised. Looks like all tokens are different on each reload. Shouldn't they be consistent during the entire session? (logged-in) https://www.mediawiki.org/w/api.php?action=query&meta=tokens&type=*