[07:55:25] 10serviceops, 10MediaWiki-General-or-Unknown, 10Operations, 10Core Platform Team (PHP7 (TEC4)), and 4 others: Some pages will become completely unreachable after PHP7 update due to Unicode changes - https://phabricator.wikimedia.org/T219279 (10Joe) >>! In T219279#5138236, @Anomie wrote: >>>! In T219279#509... [14:30:19] fsero: o/ [14:30:22] 10serviceops, 10Analytics, 10Analytics-EventLogging, 10Analytics-Kanban, and 4 others: Modern Event Platform: Deploy instance of EventGate service that produces events to kafka main - https://phabricator.wikimedia.org/T218346 (10Ottomata) Hm, I think I mostly get what you are proposing, except I'm not sure... [14:30:31] ottomata: \o [14:30:34] responded ^ :) [14:32:37] what im proposing is that you duplicate the namespace name in the release name, you don't need .release.namespace and use scap-helm to create a release named 'analytics-production' from the chart 'eventgate' so you can use wmf.releasename that will expand to 'eventage-analytics-production' [14:32:54] but i need 'eventgate-analytics' [14:32:56] so only two changes would be required, use wmf.releasename instead of wmf.chartname [14:34:02] right now, the current chart is named 'eventgate-analytics'. There are usages of .Chart.Name in the templates [14:34:20] e.g. for service-runner name [14:34:44] so I need something that will render 'eventgate-analytics', since I am changing .Chart.Name to just 'eventgate' [14:38:10] you need two things, change the occurences of wmf.chartname to wmf.releasename [14:38:13] int he chart [14:38:42] and change the release naming from production to analytics, wmf.releasename will be expanded to eventgate-analytics if the chart is named eventgate [14:38:43] but in your proposal, wmf.releasename will evaluate to 'eventgate-analytics-production' [14:39:00] because the release could be named as analutics-production [14:39:19] wmf.releasename is Chart.Name + Release.Name [14:39:23] and in fact it would help to differentiate between staging and prod cluster [14:39:29] yes [14:39:36] right, but what should I usse to get JUST 'eventgate-analytics' [14:39:36] NOT [14:39:41] 'eventgate-analytics-production' [14:40:03] ? [14:40:34] CLUSTER=staging scap-helm eventgate-analytics install -n analytics -f eventgate-analytics-staging-values.yaml stable/eventgate [14:40:52] right....? [14:40:56] will deploy in eventgate-analytics namespace in the staging cluster the chart eventgate named as analytics [14:41:13] when wmf.releasename gets expanded will be 'eventgate-analytics' [14:41:16] oh -n analytics [14:41:24] not -n analytics-staging [14:41:24] ? [14:41:46] your phab comment says -n analytics-stagaing [14:42:28] because i think is wise to also include the cluster [14:42:34] because s easier when you are kubectl get pods [14:42:43] but is not a requirement [14:42:46] just a suggestion [14:42:50] but [14:42:53] hah [14:43:18] i'm not sure what you are proposing. If you want me to use wmf.releasename, t hen I can't include the cluster in the --name , riight? [14:43:54] what im proposing is that you name the release which the information it makes the most sense to you [14:44:06] up untl now all the releases are named as production or staging [14:44:12] so since I need exactly 'eventgate-analytics', and I shouldn't use .Release.Namespace, then I should always do --name analytics or --name main (without cluster in release name), so that wmf.releasename evals to eventgate-analytics [14:45:13] hm ok, i think this makes sense. [14:45:14] one q tho [14:45:24] in one of alex's comments [14:45:29] he mentioned e.g. [14:45:33] * releases: production, staging, [14:45:51] would a possible canary be in a different CLUSTER? [14:46:29] if not, then it might be difficult to support whatever is without having a more flexible --name? [14:47:33] canary means another deployment done as a canary [14:47:45] so it would be eventgate-analytics and eventgate-analytics-test1 [14:47:59] analytics-test1 would be the name of the release [14:48:01] oh ok, so e.g. scap-helm eventgate-analytics-test1 [14:48:02] ? [14:48:06] exaclty [14:48:08] k [14:48:10] ok [14:48:11] also [14:48:26] if i change this, then the name of the monitoring sidecar and configmap volume stuff will have to change [14:48:30] but i guess that's ok [14:48:48] those names don't make a lot of sense to me right now, e.g. 'production-metrics-exporter' [14:49:15] ok i will make a patch... :) [14:49:16] thanks [14:52:28] ottomata: take in account that current policy is 1 service per namespace so you would need to do scap-helm eventgate-main -n main... stable/eventgate and another scap-helm eventgate-analytics -n analytics... stable/eventgate [14:52:57] for each cluster, which is dealing with a lot of manual steps, luckily we would have helmfle soon [14:53:13] yes that makes sens [14:53:14] e [14:53:19] that's not that different than I do now [14:53:25] except that the chart name just happens to match the service name atm. [14:53:34] and production/staging for --name [14:54:03] ok [14:56:02] 10serviceops, 10Operations, 10RESTBase, 10RESTBase-API, and 4 others: Make RESTBase spec standard compliant and switch to OpenAPI 3.0 - https://phabricator.wikimedia.org/T218218 (10CCicalese_WMF) [14:56:21] 10serviceops, 10Operations, 10RESTBase, 10RESTBase-API, and 4 others: Make RESTBase spec standard compliant and switch to OpenAPI 3.0 - https://phabricator.wikimedia.org/T218218 (10CCicalese_WMF) [15:22:27] <_joe_> fsero, mutante do you want to have the meeting? [15:22:32] <_joe_> it's just the three of us [15:22:47] <_joe_> I'm working on something right now, so I'd love the added time [15:23:04] <_joe_> I also have nothing to share, being just back from vacations [15:23:14] let's fill up at least the ehterpad so we can put some notes on the SRe one [15:23:22] <_joe_> sure [17:32:34] i was on that meeting by myself but after 5 minutes i realized it probably wouldn't happen [17:33:03] i added stuff to our section before the weekend, fwiw [17:40:02] i figured you would be just back from vacation so that's why i did that [17:46:54] 10serviceops, 10Analytics, 10Analytics-EventLogging, 10Analytics-Kanban, and 4 others: Modern Event Platform: Deploy instance of EventGate service that produces events to kafka main - https://phabricator.wikimedia.org/T218346 (10Ottomata) Talked with Fabian in IRC. The idea of using wmf.releasename (which... [18:19:20] 10serviceops, 10Operations, 10cloud-services-team: Change /r/p/ to /r/ on all hosts (where https://gerrit.wikimedia.org/r/p/ exists) - https://phabricator.wikimedia.org/T222093 (10Paladox) [18:26:34] 10serviceops, 10Operations, 10cloud-services-team: Change /r/p/ to /r/ on all hosts (where https://gerrit.wikimedia.org/r/p/ exists) - https://phabricator.wikimedia.org/T222093 (10bd808) @Paladox can you make the description more concrete? I think you have some idea from your Puppet patches about specific gi... [18:26:54] 10serviceops, 10Operations, 10cloud-services-team: Change /r/p/ to /r/ on all hosts (where https://gerrit.wikimedia.org/r/p/ exists) - https://phabricator.wikimedia.org/T222093 (10bd808) [20:32:32] 10serviceops, 10Operations, 10Thumbor, 10Patch-For-Review, 10Performance-Team (Radar): Build Thumbor packages for buster - https://phabricator.wikimedia.org/T221562 (10kchapman) [21:09:56] FYI people in this channel: for the time being don't load too much history in the Kafka dashboard https://phabricator.wikimedia.org/T222112 [22:15:45] 10serviceops, 10Operations, 10cloud-services-team: Change /r/p/ to /r/ on all hosts (where https://gerrit.wikimedia.org/r/p/ exists) - https://phabricator.wikimedia.org/T222093 (10Paladox) p:05Triage→03High [22:18:43] 10serviceops, 10Operations, 10cloud-services-team: Change /r/p/ to /r/ on all hosts (where https://gerrit.wikimedia.org/r/p/ exists) - https://phabricator.wikimedia.org/T222093 (10Paladox) [23:08:02] 10serviceops, 10Analytics, 10EventBus, 10Operations, and 4 others: Enabling api-request eventgate to group1 caused minor service disruptions - https://phabricator.wikimedia.org/T218255 (10mobrovac) [23:20:57] 10serviceops, 10Beta-Cluster-Infrastructure, 10SDC General, 10Wikidata, 10Services (done): No jobs running on beta cluster - https://phabricator.wikimedia.org/T215339 (10MMiller_WMF)