[10:42:46] errand+lunch [11:41:10] ebernhardson the struggle is real. I'm currently back to two monitors for similar reasons :D [14:12:29] random thought: The difference in opinion of the usefullness of LLMs at programming might closer represent the difference in programming capability of the observer. Perhaps people simply compare the LLM to themselves, and some people find the LLM comes up much shorter [14:15:23] I'm not much of a programmer, but I find LLMs useful for "how do you do X in this language" type questions [14:30:16] looks like the wdqs categories exporter is not actually exporting some metrics: https://phabricator.wikimedia.org/T374916#10411691 ...I guess that might be the JMX config? Checking [14:36:29] inflatador: is it because you need to ship https://gerrit.wikimedia.org/r/c/operations/puppet/+/1073529 ? [14:38:25] blazegraph_lastupdated is exported via this python exporter not JMX [14:53:22] dcausse that makes sense, thanks. Will merge now [15:07:13] sorry I'm late to pairing, got logged out [16:59:43] heading out [20:45:25] quick break, back in ~20 [21:13:15] back [21:49:14] I'm having a helluva time writing unit tests for categories lag > 2d alerts, any suggestions? https://gerrit.wikimedia.org/r/c/operations/alerts/+/1105451 [22:14:16] inflatador: with getting the metric, or with jenkins -1? [22:14:56] if you go one level deeper in jenkins you can at least find the failure output https://integration.wikimedia.org/ci/job/alerts-pipeline-test/2190/console#:~:text=AssertionError%3A%20promtool%20test%20rules%20failed%3A%20Unit%20Testing%3A%20%20blazegraph_test.yaml [22:16:45] you can also get the error locally with: docker run -it --rm --entrypoint /bin/promtool -v $PWD:/prometheus:ro prom/prometheus test rules blazegraph_test.yaml [22:16:47] there are some instructions for running locally too: https://gerrit.wikimedia.org/r/plugins/gitiles/operations/alerts/+/refs/heads/master/README.md [22:17:22] ah or that, that's even simpler :D [22:17:50] sometimes docker actually makes things simpler :) [22:28:27] looking at the thing...still not sure what you have to change to make the test pass :P Will get there though [22:37:13] inflatador: found your problem, time in unit tests starts at 0 [22:37:40] as in, jan 1 1970. So you are trying to get (time()-0) > 2880 [22:39:42] but i don't understand how to fix :S [22:42:09] inflatador: total hack, but if you set blazegraph_lastupdated to `values: "-1000000x10"` it will pass. Kinda weird though :P [22:42:59] i guess in theory you can use -172800, which is the number of seconds the test wants lag to be