[10:18:30] lunch [13:00:26] o/ [14:04:01] \o [14:05:13] o/ [14:05:45] hmm, do we need -ssd cluster for anything? Thinking should undeploy it [14:25:21] .o/ [15:32:26] heading out, have a nice week end [16:43:55] .o/ [17:05:45] was curious why the cache numbers i got in spark were higher than what our dashboard reports, turns out we were double counting misses in the dashboard. We counted miss and set in the denominator, but a miss is almost always folled by a set [17:06:05] s/folled/followed [19:01:49] I'm looking at migrating https://gerrit.wikimedia.org/r/plugins/gitiles/operations/puppet/+/refs/heads/production/modules/icinga/manifests/monitor/elasticsearch/base_checks.pp#36 to Prometheus, do we get any value of this alert? It sounds like it goes off if a specific shard gets too big? [19:12:18] looking, seems it's from 2018. I can't remember it firing, even when we did have disk issues with commonswiki_file (but perhaps we did and i'm forgetting). [19:12:21] inflatador: i think it can go away [19:12:54] ebernhardson thanks. It looks like there is a way to get the same info from our prometheus exporter by adding `--es.shards` to its startup args, but I'd just as soon do away with it if we don't get anything from it [19:17:17] noted in T431304 just in case we decide to re-enable [19:17:18] T431304: Migrate Cirrussearch checks off of icinga - https://phabricator.wikimedia.org/T431304