[08:15:49] o/ [08:31:50] o/ [11:51:16] cirrus-streaming-updater ci is having a bad day [11:51:18] https://gitlab.wikimedia.org/repos/search-platform/cirrus-streaming-updater/-/jobs/436965 [11:51:31] 3.231 mkdir: cannot create directory ‘asprof’: No space left on device [11:51:47] :/ [11:53:07] not we can do much? I can drop this asprofiler we don't use it but unsure that's going to help much [11:53:14] s/not/not sure/ [11:53:39] i think we might have to wait a bit and see if we get better resource allocation [11:53:52] sure [11:54:03] I'll hit the retry button this afternoon [11:54:16] however, I tried to restart the failing task.. and it failed again (out of disk) even on another runnier [11:55:33] :( [11:57:06] need to pick-up my son and have lunch, hopefully this gets better afternoon [11:58:38] dcausse ack. See you later. [12:55:51] lunch [13:50:58] dcausse cirrus-streaming-updater's ci fixed itself [14:01:24] \o/ [14:13:50] o/ [14:15:00] ^^ if y'all are running out of disk space on runners in the future, it might be a good idea to ask for helm in releng IRC, they are pretty responsive [14:26:59] o/ [14:27:59] dcausse do you know if all the airflow concerns here have been address? Just trying to clean up tickets https://phabricator.wikimedia.org/T380621#10448953 [14:28:21] see also https://docs.google.com/document/d/18YXTuVMAiaVwmxOktcIENrPCV_hHA_k4HEWdV2EWxvM/edit?tab=t.0#heading=h.ubfayg1k41mk [14:32:24] inflatador: yes all issues are resolved, I still need to look into the data cleanups with Gabriele to re-enable it fully [14:34:06] dcausse ACK, thanks for taking a look! [15:25:41] dcausse when we build the new opensearch plugins pkgs, will we need to each individual plugin's repo as well, such as https://gitlab.wikimedia.org/repos/search-platform/opensearch-analysis-hebrew ? [15:26:07] inflatador: yes [15:27:56] dcausse ACK, just making sure. I'll start making patches based on Erik's work in https://phabricator.wikimedia.org/T372769 . No hurry on this, just trying to scope out the work [15:28:09] sure [15:42:28] need to pick up the other kid, might be late for retro [16:04:01] retro in https://meet.google.com/eki-rafx-cxi cc ryankemper, gmodena, Trey314159 [16:07:19] looks like only Brian and myself are available today (with David maybe later), so we're canceling the retro. We'll talk in BCN! [16:07:57] gehel welp. Sorry! This one was not in my calendar [16:08:32] just finished a call, I'm aronud now if you'd like. Otherwise BCN is fine too :) [16:08:41] Oh, it looks like I forgot to invite you. [16:08:57] Nah, let's cancel for today. We'll have other opportunities. [16:09:11] just came back, +1 to cancel [16:09:39] gehel: we have a meeting with scholia today, will you be able to make it? [16:09:42] gehel np. Got the invite now [16:10:30] dcausse: yes, I'll be there. Thanks for the reminder! [16:10:37] cool, thanks! [16:26:09] https://phabricator.wikimedia.org/T370153 [16:26:40] ^ might be relevant to us (Move kafka-mirror Prometheus-based alerts from Icinga to alerts.git) [16:47:51] workout, back in 40 [17:27:08] back [17:30:59] for updating maven-hosted plugins, is the process to bump the version number in the XML files ( https://gerrit.wikimedia.org/r/plugins/gitiles/search/highlighter/+/40da2efbcad679f3b7e0bfea2461283a709240a2 ) and then run the corresponding Jenkins job, such as "search-highlighter-maven-java11"? [17:54:56] inflatador: no, if we haven't moved this one to push to gitlab you have to be a member of sonatype and be a member of the org.wikimedia.search project and build it on your laptop [17:56:12] dcausse ACK, I eventually RTFM and found it here ;( https://wikitech.wikimedia.org/wiki/Search/Elasticsearch_Administration#Deploying_plugins [19:01:21] lunch, back in ~40 [20:09:05] back [22:42:49] looks like we aren't scraping metrics from wdqs-categories JMX exporter (port 9103)...at least, I can't find them in Grafana explorer. The exporter itself seems OK [23:52:25] making some progress on the categories reload job, check it out if interested https://gitlab.wikimedia.org/repos/search-platform/sre/wdqs-categories-helm/-/blob/main/README.md?ref_type=heads [23:53:47] It'd probably work well as an airflow job if we can get Kubernetes PodOperator to work