[08:17:48] <godog>	 I missed the messages here yesterday morning, did you figure out the long catalog compilation times volans jbond ?
[08:18:25] <godog>	 also +1 to export more timing stats
[09:57:32] <jbond>	 godog: sorry had to pick up mel from the airpoit in the earlyhourse this morning so a bit of a late start :)
[09:58:40] <jbond>	 in relation to the slowness i have a change https://gerrit.wikimedia.org/r/c/operations/puppet/+/817221 which implments the recomendations from PUP-9561 (i.e. MODULES-8760) however so far i dont see much of a change
[09:58:55] <jbond>	 ill have a bit more of a play today and worse case muve the reduce to ruby
[09:59:23] <godog>	 jbond: sure, no rush! thanks for the update and the work though
[09:59:39] <jbond>	 np
[10:11:08] <godog>	 I don't have data to back this up but IIRC a significant chunk of the catalog compilation was going into querying puppetdb, even before resource import/export
[11:36:08] <jbond>	 godog: thats definetly possible.  let me try a ruby implmentation for the import functions which should gives us a much better idea if that is whats causing the problem
[11:53:09] <godog>	 jbond: SGTM!
[17:30:56] <jbond>	 godog: fyi i did a change set starting at https://gerrit.wikimedia.org/r/c/operations/puppet/+/817221 which refactors many of the rduce uses cases i see on premetheus however my testing indicates that it has ~0 difference on the compile time for promethous.  That said i think the patches are stkill worth considering as i think they are still optimisations given PUP-9561.  i thik that there 
[17:31:02] <jbond>	 is just something elses causing a ~2min lag
[17:51:05] <mutante>	 Error while evaluating a Resource Statement, Unknown resource type: 'monitoring::alerts::traffic_drop' (file: /etc/puppet/modules/profile/manifests/prometheus/alerts.pp, line: 195, column: 5) on node alert1001.wikimedia.org
[17:51:10] <mutante>	 puppet fails on alert*
[18:19:42] <mutante>	 it's being worked on over in -sre