[00:04:35] [1/3] > real 1m12.726s [00:04:35] [2/3] > user 0m0.387s [00:04:35] [3/3] > sys 0m0.208s [00:04:35] vs [00:04:42] [1/3] > real 0m0.844s [00:04:42] [2/3] > user 0m0.276s [00:04:42] [3/3] > sys 0m0.147s [00:04:48] with https://github.com/miraheze/puppet/pull/3501 [00:07:25] @rhinosf1 [00:13:48] [1/2] yes i do. [00:13:49] [2/2] it wasn't just me. another user posted about the issue on CN. https://meta.miraheze.org/wiki/Community_noticeboard#Continually_being_logged_out_while_working_on_a_page [00:13:56] hmm [00:14:37] i'm pretty chill about a few occurrences, but today was just painful for that couple hours [00:14:41] @orduin see dms pls. [00:19:47] oh memcache OOM'd [00:22:49] yikes! [00:23:14] btw thanks for what y'all do behind the scenes that most of us have no clue about. [00:23:34] tuned memcache now [00:34:13] my change to the script has massively improved things it seems [00:34:47] lowered cpu by quite a lot, lowered ram usage and improved the speed that it completed without causing much impact on the rest. [00:35:11] need to add a lock in it tho to prevent multiple scripts running if it ever manages to take longer [00:36:55] nicely done! [01:15:12] woo! [01:15:17] Nice job paladox [08:28:29] I enabled warmup for all wikis overnight and doesn’t look like we exploded [08:28:48] < 1.5k jobs [10:09:25] @paladox that's not awful [10:09:30] Having a backlog isn't too bad [10:09:37] As long as it's stable or going down [11:11:03] @paladox looks to be doing good [11:11:08] LocalGlobalUserPageCacheUpdateJob is weird though [11:11:15] it's sticking with failed jobs [11:12:05] can you do KEYS  LocalGlobalUserPageCacheUpdateJob from redis cli [11:12:14] without the space to stop discord making italics [11:14:57] I’m mobile [11:17:10] @paladox whenever time [12:01:10] that's happy news 🥳👍