[09:04:41] herron: looks like thanos-rule on titan2001 never got reloaded [09:10:17] sigh [09:10:52] Maybe we could add a notify in puppet so when pyrra reloads it will signal thanos-rule as well? [09:23:41] reloaded thanos-rule on 2001 [09:26:32] yes that's what needs to happen, for some reason I thought that was the case already [09:31:37] or we could add a dependency in the Pyrra filesystem's systemd unit [09:33:36] mmm no I had in mind PartOf/Requires but it would need to be on the thanos side [09:55:56] I'll file a tracking task [09:57:45] https://phabricator.wikimedia.org/T353691 [10:01:01] <3 [11:27:26] after the reload I still see the alerts firing :( [11:27:41] (ErrorBudgetBurn) firing: - https://alerts.wikimedia.org/?q=alertname%3DErrorBudgetBurn [12:00:32] ack, I'll take a look [12:03:16] looks like /etc/pyrra on titan2001 is not getting updated? e.g. /etc/pyrra/config/varnish.yaml is on titan1001 but isn't present on titan2001, equally /etc/pyrra/output-rules/ on titan2001 is not updated and still contains liftwing rules [14:40:05] doh! thanks [14:42:09] godog heads up that I'm merging https://gerrit.wikimedia.org/r/c/operations/puppet/+/983948 [14:42:51] yes the reload handling could be improved, thing is the operator will detect changed files and create new output rules without being reloaded itself. I was thinking something like PathModified on output-rules to issue a reload, until upsteram adds a custom reload command/url feature [14:45:44] inflatador: +1'd [14:53:25] godog excellent..merged. Working on a patch to use the new param ;) [15:36:35] since when did the CI pipeline for alerts get so useless? https://integration.wikimedia.org/ci/job/trigger-alerts-pipeline-test/1483/console :D [15:37:29] you have to follow the links [15:37:33] to https://integration.wikimedia.org/ci/job/alerts-pipeline-test/1485/console [15:37:41] not sure why :D [15:38:46] that's very unintuitive...I clicked all tree links (two are the same) and none of those led to the output [15:39:04] one has to click *again* to get the output of the job [15:39:34] IIRC that used to work "properly" in the past [15:41:41] I'm not saying it wasn't :D just how to get there from it [15:41:50] probably some parent/child job relation? [15:42:02] yeah, sure - did not mean to sound harsh. sorry [15:42:15] you weren't ;) [15:42:47] it has 2 jobs from operation-misc.yaml: [15:42:47] - 'trigger-{name}-pipeline-{pipeline}' [15:42:53] - '{name}-pipeline-{pipeline}' [15:43:22] possibly related