[08:53:39] For the first test planned for week of Oct 19th, discussing with jayme yesterday we may want to do it early in the week, to give us time to fix things (eg memcached related) before the beginning of the DC switchover [09:53:21] <_joe_> tbh if we fix the things we /assume/ caused the capacity issues, we should test switching mw-web well before [11:15:25] the last fix in Central Notice (T439239) is progressing, we can check with Eliott when he thinks it will get merged, and potentially test as early as next week [15:59:22] cdanis: last week you mentioned in this channel having "metrics for # of PHP workers that are in post-send" - was a task created for it which we should request MW Eng to prioritize? do you feel this is a must have for next DC Switchover? [16:04:16] not sure if a task was created, I'd have to look [16:04:23] it's not a must-have [16:04:44] and exporting a gauge from mediawiki is tricky heh, I think we first should ask for a latency histogram [19:42:25] Is it okay for me to do the first deploy of a chart to production on my own (editcheck-headless, which has been trialed in staging for a while), or do you prefer someone serviceops-related to be keeping an eye on that? [19:49:54] Kemayo: without knowing the details of the specific service, as long as serviceops has taken care of things like namespace creation, etc. in the production k8s clusters, then I think service owners are generally encouraged to drive the later stages of [0] (if I'm understanding the question correctly). [19:49:54] in any case, I'm here in case you run into anything odd :) [19:49:54] [0] https://wikitech.wikimedia.org/wiki/Kubernetes/Add_a_new_service#Deploy_the_service [19:52:00] swfrench-wmf: Thanks! I largely wanted to make sure that first-deploys weren't a thing that warranted any special oversight. [19:52:18] I'll go ahead and shout if anything goes wrong. [19:53:09] * swfrench-wmf thumbs up [19:53:33] as long as the process has worked, it should be just about as exciting as a normal `helmfile apply` [19:56:32] All seems to have gone smoothly, indeed. [19:57:40] \o/ [19:57:59] I see pods :) [19:58:01] awesome