[06:00:01] Hello, I'm about to reboot the turnilo server an-tool1007.eqiad.wmnet between 06:15 UTC and 06:30 UTC - please let me know if I should wait [06:05:13] <_joe_> steve_munene: go on, this is probably the only time of the day no one's using it :) [06:06:06] Great [07:39:49] <_joe_> uhm sirenbot? [07:40:08] <_joe_> ah sorry somehow I missed the updated topic [11:16:40] moritzm: good to merge `Muehlenhoff: postgres: Remove support for Stretch (51cbf39d39)` ? [11:20:52] arturo: sorry, yes please go ahead [11:21:02] um, I thought we had at least one postgres cluster (osmdb, clouddb1003/4.clouddb-services) on stretch in wmcs [11:21:04] np, merging now [11:32:54] there was wikilabels, but that got upgraded: https://phabricator.wikimedia.org/T307389 [13:28:17] jbond pcc is broken? https://puppet-compiler.wmflabs.org/pcc-worker1003/38168/cp3050.esams.wmnet/prod.cp3050.esams.wmnet.err [13:29:06] vgutierrez: looking [13:44:42] vgutierrez: sorry about that should be good now https://puppet-compiler.wmflabs.org/pcc-worker1002/38171/cp3050.esams.wmnet/index.html [13:45:01] cheers <3 [13:45:05] np [16:02:34] cdanis, XioNoX - still very WIP, we need to check 1000 things, but: https://w.wiki/5x$Y :) [16:03:21] niiiice! [16:03:25] amazing [16:04:21] elukey: nice!!! [16:04:48] special thanks also to godog for the work :) [16:05:57] woot woot, excellent [16:06:16] the idea for the moment is to keep sampled_live separated and set a retention of say 12h [16:07:48] the other useful thing is that now there's a webrequest_sampled kafka topic to be looked at, which is directly ingested into druid/turnilo above [16:10:20] <_joe_> that's great [16:10:40] <_joe_> volans: please adapt your script to aggregate from there :P [16:12:23] <_joe_> cdanis: maybe we should add https://w.wiki/5x$c somewhere in the ddos response docs [16:13:25] _joe_ some basic sanity checks are still needed, let's keep it in mind [16:13:38] <_joe_> elukey: wdym? [16:13:54] <_joe_> I'm really excited, this work is really amazing folks [16:13:55] _joe_: I'm in a meeting talking about this ;) [16:14:07] <_joe_> did you use benthos for this? [16:14:08] _joe_ comparing the sampling for example, DE uses hive buckets meanwhile we use the varnishkafka sequence id etc.. [16:14:15] yep Benthos [16:14:19] <_joe_> elukey: w/e [16:14:25] it runs on centrallog nodes for the moment [16:14:27] <_joe_> it's good enough I'd say :) [16:14:33] elukey: and we have some of the refinements but not all? [16:14:44] <_joe_> elukey: I have another application for benthos soon :) [16:14:47] I see ASN, ISP, country code, the x-analytics subfields, but not is_pageview [16:15:04] cdanis: the only missing thing is "is_pageview" but the rest should be equal to webrequest_sampled_128 [16:15:08] nice [16:15:39] we set it to "-" by default, I didn't check yet how it is calculated [16:15:49] I think it's a UDF in Java [16:15:50] is_from_public_cloud and requestctl should work though [16:15:57] so probably can't do is_pageview, SRE doesn't usually use that one though [16:16:44] elukey: awesome news :) [16:17:14] yeah.. for the use case of webrequest_sampled_live is_pageview isn't a must IMHO [16:17:20] <_joe_> yeah it's one dimension we don't need *live* [16:18:02] <_joe_> from the task I got the feeling we weren't doing much progress, this is an amazing gift elukey godog <3 [16:18:08] yeah indeed [16:18:42] it looks like we need to overflow their beer counter [16:19:14] _joe_: you are welcome <3 [16:19:22] vgutierrez: haha! bring it on, I'm ready [16:26:17] elukey: awesome <3 [16:26:31] hmmm [16:27:06] my chromium is complaining about the Permissions-Policy header: "unrecognized origin: 'intake-analytics.wikimedia.org'" [16:32:18] great job elukey, godog [16:33:25] thank you claime, appreciate it [17:27:30] jbond: something is off with pcc https://integration.wikimedia.org/ci/job/operations-puppet-catalog-compiler/38190/console [17:27:56] pcc says that results are available at https://puppet-compiler.wmflabs.org//38190/ [17:28:28] but the URL is missing the pcc-worker directory [17:29:06] expected URL is https://puppet-compiler.wmflabs.org/pcc-worker1003/38190/ [17:29:51] vgutierrez: sorry just made a bade change https://puppet-compiler.wmflabs.org/output/38190/ should also work [17:39:24] vgutierrez: shuld be fixed now, any reports that where wrong you can just add `output` in the path before the job id [17:40:14] fyi i plan to do a small release soon, i have been using it fr the past week so shouldn;t break any thing but pleas eping if you see issues [17:40:18] thanks [17:42:52] np