[04:08:19] <musikanimal>	 is there anything that can be done about the replag?
[05:42:06] <silky>	 Hello Andrew, I'm an outreachy intern and I've requested for membership in the Toolforge project to work on the APIs for Leaderboard.  Link to my request - https://toolsadmin.wikimedia.org/tools/membership/status/533 . I've updated the comment with details of the APIs. Can you please help me in this regard. Thank you! 
[06:28:34] <zhuyifei1999_>	 silky: I approved it
[08:20:55] <andrewbogott>	 thanks zhuyifei1999_ 
[13:55:02] <zhuyifei1999_>	 (np)
[18:27:58] <Wurgl>	 I am wndering why I always connect to labsdb1011 when I type "sql de"? Should'nt I sometimes connect to labsdb1009 or labsdb1010?
[18:29:55] <Reedy>	 luck?
[18:30:29] <Wurgl>	 bad luck … because labsdb1011 seems to be slower
[18:33:19] <Reedy>	 you can try a different cluster too
[18:34:30] <Wurgl>	 If I knew how … connecting to labsdb1010.eqiad.wmnet does not work
[18:36:11] <Reedy>	 if you use `sql --cluster=web dewiki` that'll put you on labsdb1010
[18:46:55] <Wurgl>	 Was there a change in the configuration? I am very sure that dewiki.analytics.db.svc.eqiad.wmflabs connected to labsdb1010 a few times. Somewhen end of May the behaviour seems to be different
[18:57:57] <Krenair>	 Wurgl, there may have been, I don't think the exact routing is at all a stable API
[19:11:16] <Wurgl>	 https://wikitech.wikimedia.org/wiki/Help:Toolforge/Database <-- I read here labsdb1011 is "potentially slower". Well, instead of 2 Minutes 37 seconds some bunch of queries take 9 minutes 44 seconds. Which is not bad when there is just one, but for the overall job it is less than 3 hours vs. close to 12 hours
[19:20:13] <Reedy>	 Is the db maintenance still ongoing?
[19:20:16] <Reedy>	 So hence "known issue"
[22:01:52] <bstorm_>	 There was a change made by the DBAs in how the replicas are arranged
[22:02:05] <bstorm_>	 labsdb1011 is now the analytics replica alone.
[22:02:11] <bstorm_>	 The web replicas are now a roundrobin