[00:11:55] <swfrench-wmf>	 just refreshed my memory of how the report consumed by the `PyBal backends health check` is formed, and tl;dr - it's expected that the check would not fail in this case :)
[00:11:56] <swfrench-wmf>	 longer version:
[00:11:56] <swfrench-wmf>	 * my (incorrect) recollection was that a persistently down-but-enabled host (note, "enabled" in pybal means "pooled" in the confctl sense) was sufficient to trigger this
[00:11:56] <swfrench-wmf>	 * however, the report is looking at down-but-pooled (note, "pooled" in pybal means active for load balancing), which this was clearly not (e.g., the failing host was internally depooled, as the service remained above the depool threshold)
[08:37:18] <moritzm>	 if anyone in SRE has ever used dstat on our servers: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1145100
[08:57:45] <akosiaris>	 dstat, wow, that brings back memories
[09:11:51] <moritzm>	 depending how fond those memories are, it seems the upstream maintenance is also up for grabs :-)
[11:18:13] <_joe_>	 I haven't used dstat in 8 years I think
[12:06:29] <klausman>	 https://www.youtube.com/watch?v=EJizB6hFON8 Interesting talk on how Booking moved from Blade servers to VM-alike partitioning of hosts
[12:10:24] * Emperor has "fond" memories of blade servers from the Sanger
[12:11:52] <marostegui>	 klausman: I worked at booking.com and only handling the purchases of them was a nightmare
[12:18:28] <klausman>	 I can imagine (but sdon't wanna :))
[12:19:22] <moritzm>	 "Only five blade servers left at this price on our site"
[12:21:30] <Emperor>	 lol
[12:21:55] <Emperor>	 marostegui: did you know Andrew Mobbs by any chance? He was a DBA for booking.com for a bit
[12:22:26] <marostegui>	 Emperor: Nope, he wasn't in the team when I joined/left, do you know which years he was there?
[12:22:52] <Emperor>	 not beyond "donkeys years ago", I'm afraid.
[12:22:57] <marostegui>	 haha
[12:44:28] <XioNoX>	 bblack, marostegui, arnoldokoth, volans, fyi, I'm in the process of upgrading cr3-eqsin, eqsin is depooled
[12:44:58] <volans>	 XioNoX: ack, thanks for the heads up, anything to expect? need a hand for silencing alerts?
[12:45:53] <XioNoX>	 volans: I downtimed cr3-eqsin, there will be some alerts for bgp sessions going down for example, so maybe a bit of noise but nothing more
[12:46:06] <volans>	 ack
[12:49:01] <bblack>	 ack :)
[13:13:51] <inflatador>	 interesting, thanks for sharing klausman 
[13:19:36] <arnoldokoth>	 XioNoX: Ack.
[13:33:17] <volans>	 XioNoX: let us know once completed, so far so good :)
[13:33:26] <XioNoX>	 volans: completed :)
[13:33:32] <volans>	 nice, thx
[13:36:06] <XioNoX>	 I'll wait a bit before repooling just in case, but it can be done anytime now both routers are 100% operational
[13:41:23] <volans>	 ack
[13:47:20] <XioNoX>	 volans: and oncall, I'm repooling eqsin
[13:47:28] <volans>	 great, thx
[13:55:04] <godog>	 inflatador: when you get a moment, an easy one for your eyes https://gerrit.wikimedia.org/r/c/operations/puppet/+/1145138
[13:59:05] <inflatador>	 godog 👀
[14:01:48] <inflatador>	 +1'd, thanks for spotting this
[14:02:13] <hnowlan>	 general heads-up: we're just finalising the migration of mobileapps/pcs APIs out of restbase at the moment - we've been doing it in chunks and the increase in load hasn't been anything alarming so far but this will be a bit of a jump 
[14:03:19] <godog>	 inflatador: sure no worries, cheers for taking a look
[15:55:59] <btullis>	 I still use dstat 😢
[16:40:14] <swfrench-wmf>	 brett: is it alright if I pick up your admin patch?
[16:40:34] <brett>	 yep!
[16:41:04] <swfrench-wmf>	 brett: ah, it seems you have the lock :)
[16:41:18] <swfrench-wmf>	 you're good to go on mine
[16:41:36] <brett>	 oh, derp, sorry!
[16:41:40] <brett>	 doing now
[16:41:49] <swfrench-wmf>	 thanks!