[05:24:50] (03CR) 10Krinkle: [C: 04-1] "Minor but important nit: We should use the MediaWiki.* namespace for metrics send from outside MediaWiki. This namespace is prepended by M" [analytics/refinery] - 10https://gerrit.wikimedia.org/r/392703 (https://phabricator.wikimedia.org/T176785) (owner: 10Joal) [05:25:16] (03CR) 10Krinkle: [C: 04-1] "Correction: s/We should use/We should not use/" [analytics/refinery] - 10https://gerrit.wikimedia.org/r/392703 (https://phabricator.wikimedia.org/T176785) (owner: 10Joal) [05:27:13] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3780217 (10Krinkle) >>! In T176785#3777250, @JAllemandou wrote: > The existing metric is named `restbase.requests.varnish_requests`, and we plan to go with... [08:58:33] * elukey afk for a bit! [11:34:59] the first run of the eventlogging cleaner as cron is executing now \o/ [11:42:40] Yay elukey :) [12:08:03] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3780901 (10JAllemandou) @Krinkle: I'm happy to use a less functional approach, and have for instance `analytics.varnish_requests.restbase` and `analytics.v... [12:18:48] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3780948 (10Pchelolo) > Only concern is that it changes the existing metric for restbase. @mobrovac and @Pchelolo, is that a big deal? That will not be des... [12:24:07] (03CR) 10Addshore: [C: 04-1] "*agrees with Krinkle*" [analytics/refinery] - 10https://gerrit.wikimedia.org/r/392703 (https://phabricator.wikimedia.org/T176785) (owner: 10Joal) [12:51:51] * elukey coffee [12:56:36] Taking a b [12:56:41] break a-team [14:12:27] 10Analytics-Tech-community-metrics, 10Developer-Relations (Oct-Dec 2017): Advertise wikimedia.biterg.io more widely in the Wikimedia community - https://phabricator.wikimedia.org/T179820#3781299 (10Aklapper) [14:41:11] joal: whenever you are back I'd like to ask you some questions about how to restart druid daemons on druid100[123] [14:41:18] (to enable prometheus metrics) [14:54:47] 10Analytics-EventLogging, 10Analytics-Kanban: Purge refined JSON data after 90 days - https://phabricator.wikimedia.org/T181064#3781379 (10Ottomata) Naw, tbayer.popups isn’t gonna be touched. [15:04:44] 10Analytics, 10Operations, 10ops-eqiad, 10User-Elukey: Check analytics1037 power supply status - https://phabricator.wikimedia.org/T179192#3781423 (10elukey) ping :) [15:25:18] ottomata: o/ [15:27:49] Heya elukey - In da cave for ops sync [15:28:02] joining in a minute [15:31:19] coming! [15:50:19] 10Analytics-Kanban: Procure hardware to refresh jupyter notebooks - https://phabricator.wikimedia.org/T175603#3781547 (10Ottomata) a:03Ottomata [15:52:21] 10Analytics, 10Analytics-Cluster: Upgrade Analytics Cluster to Java 8 - https://phabricator.wikimedia.org/T166248#3781548 (10Ottomata) a:05Ottomata>03elukey [15:52:48] 10Analytics-Cluster, 10Analytics-Kanban, 10User-Elukey: Upgrade Analytics Cluster to Java 8 - https://phabricator.wikimedia.org/T166248#3781549 (10elukey) [15:52:50] 10Analytics-Cluster, 10Analytics-Kanban, 10User-Elukey: Upgrade Analytics Cluster to Java 8 - https://phabricator.wikimedia.org/T166248#3781550 (10Ottomata) [16:08:07] 10Analytics, 10Operations, 10ops-eqiad, 10User-Elukey: Check analytics1037 power supply status - https://phabricator.wikimedia.org/T179192#3781597 (10RobH) I think replacing bad powersupplies on out of warranty servers is likely a waste of money (as other parts will also go bad with older systems), however... [16:09:18] !log restarting eventlogging services on eventlog1001 [16:09:19] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log [16:13:30] 10Analytics, 10Analytics-Cluster, 10Spike: SPIKE: experimenting with importing Revision history from XML dumps into an easier to use format, Avro - https://phabricator.wikimedia.org/T78405#844429 (10JAllemandou) Some code exists for converting XML dump in hadoop: https://gerrit.wikimedia.org/r/#/c/361440/ It... [16:18:49] 10Analytics, 10Analytics-Cluster, 10Spike: SPIKE: experimenting with importing Revision history from XML dumps into an easier to use format, Avro - https://phabricator.wikimedia.org/T78405#3781633 (10JAllemandou) @ggellerman : I think we can close this task, just want to confirm with you. [16:20:15] 10Analytics, 10Analytics-Dashiki: Browser-major numbers represented as a percentage (when they're not) on the desktop tabular view - https://phabricator.wikimedia.org/T181164#3781638 (10Jdforrester-WMF) [16:47:04] joal: druid restarts completed! [17:02:46] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3781820 (10JAllemandou) @Krinkle and @Addshore: After looking once more in graphite, would `mw.api.varnish_requests` be ok for you guys? This would allow u... [17:06:28] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3781829 (10Addshore) As far as I am aware, for those with access to the graphite box it is fairly trivial to move and merge metrics. "analytics.varnish_re... [17:07:05] elukey: apparently puppet linter makes you do class { '::standard': } etc. for non profile classes [17:07:11] ¯\_(ツ)_/¯ [17:08:02] yeah I suspected that :( [17:24:43] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3781893 (10Anomie) I'm a bit confused here. Is this task about adding a count for the action API, turning an existing restbase-only count into a restbase+a... [17:25:39] 10Analytics, 10Analytics-Cluster, 10Spike: SPIKE: experimenting with importing Revision history from XML dumps into an easier to use format, Avro - https://phabricator.wikimedia.org/T78405#3781894 (10ggellerman) @JAllemandou Thanks for asking! However, I defer to @Ottomata on this question. [17:27:46] 10Analytics, 10Analytics-Cluster, 10Spike: SPIKE: experimenting with importing Revision history from XML dumps into an easier to use format, Avro - https://phabricator.wikimedia.org/T78405#3781924 (10Ottomata) 05Open>03Resolved a:03Ottomata [17:34:46] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3781971 (10Pchelolo) >>! In T176785#3781893, @Anomie wrote: > I'm a bit confused here. Is this task about adding a count for the action API, turning an exi... [17:35:49] ottomata: oooooh, statistics::explorer [17:36:14] shhhh thank luca for that name [17:36:21] these stat classes are sooo baad [17:36:23] very hard to work with [17:36:27] priority to fix is very low [17:36:29] yaaaarrr [17:37:14] :D [17:37:25] I was trying to find an appropriate explorer meme, but apparently they ar eall rubbish [17:37:27] good name! [17:45:54] * elukey off! [18:42:19] 10Analytics, 10WMDE-Analytics-Engineering, 10Patch-For-Review, 10User-GoranSMilovanovic: A statbox to update the WDCM system - https://phabricator.wikimedia.org/T181094#3782283 (10Ottomata) OK! I've installed R and a mysql .conf pw file on stat1004. If you are in the analytics-privatedata-users group on... [18:42:44] 10Analytics-Kanban, 10WMDE-Analytics-Engineering, 10Patch-For-Review, 10User-GoranSMilovanovic: A statbox to update the WDCM system - https://phabricator.wikimedia.org/T181094#3782284 (10Ottomata) [18:43:12] 10Analytics, 10Analytics-Cluster, 10WMDE-Analytics-Engineering: Install R on stat1004? - https://phabricator.wikimedia.org/T174890#3782286 (10Ottomata) [18:43:15] 10Analytics-Kanban, 10WMDE-Analytics-Engineering, 10Patch-For-Review, 10User-GoranSMilovanovic: A statbox to update the WDCM system - https://phabricator.wikimedia.org/T181094#3779313 (10Ottomata) [18:46:40] 10Analytics-Kanban, 10Patch-For-Review, 10Services (watching): Add action api counts to graphite-restbase job - https://phabricator.wikimedia.org/T176785#3782292 (10JAllemandou) @Pchelolo is correct, idea is to have both restbase and mw-action-api hourly varnish-requests counts in graphite. [18:48:28] A-team, with your approval, I'll take friday afternoon to go to a meetup on warp10 (http://www.warp10.io/) [18:49:05] warp10 is developped 10 minutes from home, and seems very powerfull when it comes to huge volume of timeseries [18:49:21] (OVH uses it for its metrics AFAIK) [19:08:01] 10Analytics-Kanban, 10Operations, 10hardware-requests: eqiad: (2) hardware access request for jupyter notebook refresh (SWAP) - https://phabricator.wikimedia.org/T175603#3782420 (10Ottomata) p:05Triage>03Normal [19:12:10] 10Analytics-Kanban, 10Operations, 10hardware-requests: eqiad: (2) hardware access request for jupyter notebook refresh (SWAP) - https://phabricator.wikimedia.org/T175603#3782444 (10RobH) A few questions: * Do you have a minimum clock speed you need on the CPU, or just whatever the best price point is? ** We... [19:16:18] nuria_: , yt? [19:16:28] ottomata: yessir [19:16:35] notebook hw refresh [19:16:40] am assuming we don't need much disk space [19:16:48] but maybe we do? do we want to support folks saving large datasets there? [19:20:47] ottomata: i think that will fire back as we cannot really support processing for n people for largedatasets [19:20:51] ottomata: right? [19:21:05] ottomata: so probably best to store and process large datasets in hadoop [19:21:23] agree, but the default storage if we don't ask is 200G [19:21:25] which maybe is enough? [19:21:34] just dont' want to get to a spot where people fill up the disks [19:21:46] i think we can afford larger HDDs if we wanted them [19:21:48] ottomata: how large are notebooks? let me see, i have only done a couple [19:22:34] yeah, i guess its better not to have large files there, that would encourage people to do big data processing in the notebook process itself, which is not what we want [19:22:45] ottomata: ya that will go kaput [19:22:49] joal, going to warp10 seems cool to me! :] [19:22:53] notebooks are tiniy [19:23:11] ottomata: mine is not even 500k [19:23:20] ottomata: so i'd say 200G is pelnty [19:23:23] *plenty [19:23:54] Oh by the way ottomata -- I've tested zeppelin for spark - it;s really great :) [19:24:00] oh ya? [19:24:23] ottomata: if we manage to have spark working in jupyter, no need, but if we don't, zeppelin would be good [19:24:25] ok nuria, we'll go with the 200Gs [19:24:41] 10Analytics-Kanban, 10Operations, 10hardware-requests: eqiad: (2) hardware access request for jupyter notebook refresh (SWAP) - https://phabricator.wikimedia.org/T175603#3782479 (10Ottomata) > Do you have a minimum clock speed you need on the CPU, or just whatever the best price point is? No, but as always,... [19:25:41] Thanks mforns for the +1 :) I also like the idea of learning some on that thing [19:26:30] joal: sounds maybe a little like Would this benefit from SSDs? [19:26:31] oop [19:26:34] like [19:26:35] https://www.timescale.com/ [19:27:24] nuria_: other things. superset. this one is really tricky. [19:27:41] deb package would be very hard, as there are a lot of non debianized dependencies [19:27:45] ottomata: i have it on my list of goals-to-talk, why is it tricky [19:27:52] ah sorry [19:27:53] python deployment is not easy [19:28:06] maybe i should see how halfak|Lunch is doing his ores stuff [19:28:08] and try to emulate [19:28:09] ottomata: and making it all (package and deps) [19:28:17] one debian? [19:28:19] joe has a proposal for python packaging deployment in the work [19:28:22] ottomata: from what I have read on warp10 and what I see on timescale, they are a bit different ) [19:28:31] nuria_: not really sure how do to that without virtualenv... [19:28:36] hmm [19:28:38] maybe... [19:28:47] ottomata: fat -jar style [19:28:57] yeah, but pythonpaths are weirder than nodes ones. [19:29:04] it might be possible [19:29:23] ottomata: ya python paths not so hot [19:29:40] ottomata: this sounds like a good candidate for tasking [19:29:46] hmm https://wikitech.wikimedia.org/wiki/ORES/Deployment [19:30:04] joal: i just double checked, i seem to be able to run pyspark just fine on SWAP [19:30:37] ottomata: that is a node style depployment for python right? [19:30:45] ottomata: they are deploying build source [19:30:51] ebernhardson: cooooool, connecting to yarn though? [19:30:55] ebernhardson: , can you do [19:30:58] pyspark --master yarn [19:30:58] ? [19:30:59] ebernhardson: do you mind sharing your comand line? [19:31:16] joal: no command line, just straight jupyter. one sec [19:32:14] ottomata: or maybe i am not getting it [19:32:33] ottomata: but they are deploying python like we deploy aqs [19:32:37] ottomata: eight? [19:32:42] ottomata: sorry, right? [19:32:52] joal: hmm, so https://phabricator.wikimedia.org/P6368 but i just double checked and thats running spark locally not against yarn. Can probably make it work against yarn though checking [19:33:03] (but still able to access hdfs and such) [19:35:08] nuria_: no, because virtualenv paths are hardcoded [19:35:12] you can't just move virtualenvs around [19:35:21] they use a virtualenv to create a wheel (?) and then deploy that (?) [19:35:21] i think [19:35:23] i'm going to try [19:35:38] ok [19:35:55] Who is looking at ORES config? [19:36:38] halfak: me [19:36:42] we want to deploy a python thing [19:36:50] not our code [19:36:55] just a service that we need to package and deploy [19:36:59] but there are a lot of python deps [19:37:05] so .deb package is not looking fun [19:37:11] i don't know much about wheels... [19:39:55] joal: hmm, trying to run against yarn gets stuck in the Accepted stage in yarn :S i'll have to spend more time with it to figure out why, but spark local works which might be enough depending on what people are doing. [19:40:13] aye, i think i remember there being some problem lik ethat [19:40:16] don't remember why atm though [19:40:21] Thanks ebernhardson for the check :) [19:40:37] ebernhardson: How do you try to launch spark within yarn in notebook? [19:40:46] joal: conf.setMaster('yarn-client') [19:40:52] (before creating the SparkContext) [19:40:57] Right [19:41:01] Makes sense [19:47:52] (03PS18) 10Mforns: [WIP] Add scala-spark core class and job to import data sets to Druid [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/386882 (https://phabricator.wikimedia.org/T166414) [19:48:43] ebernhardson: May I kill your non-running pusaprk-apps ? [19:48:56] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Add scala-spark core class and job to import data sets to Druid [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/386882 (https://phabricator.wikimedia.org/T166414) (owner: 10Mforns) [19:50:44] (03PS19) 10Mforns: [WIP] Add scala-spark core class and job to import data sets to Druid [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/386882 (https://phabricator.wikimedia.org/T166414) [19:50:53] wow so halfak, do you recreate a new virtualenv from deployed wheels every time you deploy? [19:50:55] just looked at [19:51:02] scap/cmd_worker.sh [19:54:40] joal: yes definately [19:55:12] ebernhardson: They dies by themselves :) [19:56:12] hmm, maybe i can make a .deb package from these wheels. [19:56:36] hmm, no, [19:56:37] hm no [20:00:20] ottomata: the virtualenv sorta gets recreated each time. You can also look at https://github.com/wikimedia/labs-striker-deploy and https://github.com/wikimedia/search-MjoLniR-deploy as other python deployments that were also based on ORES work [20:00:37] ottomata, that's right. [20:00:58] the venv recreation doesn't take that long with wheels. [20:01:01] They are pre-compiled. [20:01:09] aye [20:01:11] It's just a matter of decompressing a bunch of zip files. [20:01:20] wheel == zip with a specific structure [20:01:32] is there a reason you don't commit and deploy the virtualenv itself? if you can guaruntee that it will be deployed to the exact same path? [20:01:35] where it was built? [20:09:29] halfak: ^? [20:22:48] mforns: interesting findings after dumps computation 1 [20:23:01] ahaaaa [20:23:11] mforns: batcave for a minute? [20:23:12] what did you find? [20:23:15] sure [20:40:34] Hi Joal, Are you around? Seeking your advice on a few things... [20:40:44] Hi Shilad ! Here I am :) [20:40:50] Shilad: batcave, or IRC? [20:40:59] IRC okay, I think! [20:41:03] okey [20:41:26] I wonder Shilad which is better: okay or okey - probably a matter of accents :) [20:41:50] Hah! Q1: I am on the very last Spark job (Hooray!) that converts pageview session logs to concepts. [20:42:13] Option 1: Enhance session hive schema to store both local page ids and wikidata concepts. [20:42:53] Option 2: Mirror the language-specific schema with a duplicate where page ids are replaced by entity ids. [20:43:09] I like option 1, even though it makes the pageview processing more complex. [20:43:14] Does that seem reasonable? [20:43:34] hm - I need to think a bit [20:43:43] Better in batcave? [20:43:48] Shilad: I think that intuitively I prefer option 2, but I can't really say why [20:44:16] Shilad: possibly because it makes each of the dataset easier to use [20:44:39] Oh, ans also I think becasue you won't have concepts while building sessions, would you? [20:44:49] I could have them, I think. [20:44:58] Ah, different matter then [20:45:37] I would probably to a broadcast join, and it's a big datastructure (1.5GB by my accounting). [20:45:45] In option 1 we would have a table with 3 columns? [timestamp, page_ids, concepts], where page_ids and concepts are arrays of the same size? [20:46:52] RIght now it is: pages ARRAY [20:47:04] I would do what you are saying... add concept ids. [20:47:22] and it would be the same size. [20:48:04] Shilad: or put the concept_id as part of the struct in array, could make sense as well [20:48:24] Right! Sorry, that's what I was thinking. [20:48:30] :) [20:48:38] Ok, I think that;s a good idea :) [20:49:11] Great! Do you think a 1.5GB datastructure is too brutal to broadcast? [20:49:12] Shilad: It'll push our users of sessions to think of them in term of wikidata concepts in addition to page_ids :) [20:50:07] Shilad: I've not done that much broadcasting, so I don't have strong opinion here - I think it;s big, but if using executor settings wisely (4 cores and 16G RAM for instance, could work like a charm) [20:51:05] This is what I'm thinking too... I'm not sure about broadcast semantics. Do you know if Spark broadcasts once per partition, or once per executor? I presume it's smart enough to only do once per partition. [20:51:16] Sorry... I mean once per executor. [20:51:29] So if there is 1000 partitions but only 32 executors, it's only 32 broadcasts. [20:51:58] Shilad: I'm pretty sure it does it the way you suge [20:52:01] suggest sorry [20:52:17] Then it shouldn't be too bad. [20:52:57] FYI, I have actually built the sitelinks table: shilad.sitelinks if you want to take a look. Has page information before and after redirects, and also the wikidata entities. [20:53:26] That's really great Shilad :) [20:53:30] Okay, last question: This is a big CL! Would it be better to pull all of it out into its own project? [20:53:56] Shilad: I'm struggling for the moment with an issue that takes most of my time, but I'll be glad to help moving this to production when I'm done :) [20:54:29] Thanks. I understand. When that happens, do you still think it's best for it to stay in analytics-refinery despite the size? [20:55:07] Shilad: I like it staying in refinery, it's where all that type of code belongs [20:55:42] Shilad: I suggest having 2 packages: pageviewsessions, wikidataentities [20:55:48] does it make sense o you? [20:56:02] Also, I'm terrible with names, we could probably do better [20:57:29] Yes! I did that but called them slightly different things: pagesessions and sitelinks [20:57:43] Totally happy to change them if you prefer pageviewsessions and wikidataentities [20:57:48] do you mind adding wikidata.sitelinks to the first one ? [20:57:56] last one sorryb [20:57:59] late here ;) [20:58:55] Absolutely. And my questions are done :) Thanks for your help on this! [20:59:07] Sorry for intruding on your evening. [20:59:12] Shilad: no problem at all, thank you again ! [20:59:27] Shilad: I'm actually still on my issue taking most of my time, nothing to worry ;) [20:59:54] Hah. That does not sound fun. But maybe it is?! [21:00:07] Have a good night! [21:00:43] Shilad: it's long and complex, but not that bad - just too long :)Have a good end of day too - [21:06:51] mforns, ottomata --> IT WORKED !!!! 0.02% diff for all times by faking correct reconstruction with 2017-09 snapshot [21:07:04] wooohooo [21:07:25] Ok - Now that I have it, it's just the matter of correctly patching and ask Erik about his discrepeny [21:07:30] I'm really happy :) [21:07:40] you rock joal nice job [21:09:50] * joal goes to bed, moving like that http://reactiongifs.me/wp-content/uploads/2014/01/jimmy-fallon-elmo-happy-dance-saturday-night-live.gif [21:11:48] Bye a-team, see you tomorrow ! [21:14:18] byyye [21:14:23] joal AMAAAAING [21:17:35] (03PS1) 10Ottomata: Deploy repo for superset (python) [analytics/superset/deploy] - 10https://gerrit.wikimedia.org/r/392933 (https://phabricator.wikimedia.org/T166689) [21:20:06] fwiw i double checked why notebook1001 couldn't connect to yarn for spark, pretty sure its the same firewall that prevented 1005: https://phabricator.wikimedia.org/T170496 [21:23:35] ebernhar|lunch: pyspark working in jupyter is really a fantastic news :) [21:23:46] ebernhar|lunch: Many many thanks :) [21:24:29] joal: np! [21:28:52] ottomata: maybe if we have some time, we could test that as well: https://toree.apache.org [21:29:44] hm [21:32:17] joal jaja ! many thnaks for your work [21:32:43] nuria_: I plan to document my findings tomorrow in ticket [21:33:07] joal: now some might be in order [21:33:14] huhu [21:39:12] (03CR) 10Nuria: Spark job to create page ids viewed in each session with both language specific ids and wikidata data ids. (031 comment) [analytics/refinery/source] (nav-vectors) - 10https://gerrit.wikimedia.org/r/383761 (https://phabricator.wikimedia.org/T174796) (owner: 10Shilad Sen) [21:43:45] 10Analytics, 10Phabricator: Create phabricator space for tickets with legal restrictions - https://phabricator.wikimedia.org/T174675#3782735 (10Nuria) The space should include the following members: members of analytics team @mforns @Ottomata @Milimetric @JAllemandou @elukey @fdans plus @Nuria plus CTO: @VCole... [22:01:50] (03CR) 10Ottomata: "Eric, cause I'm new at this python wheels thing...what do you think?" [analytics/superset/deploy] - 10https://gerrit.wikimedia.org/r/392933 (https://phabricator.wikimedia.org/T166689) (owner: 10Ottomata) [22:06:36] (03CR) 10Nuria: Update cassandra load jobs to local quorum write (031 comment) [analytics/refinery] - 10https://gerrit.wikimedia.org/r/392624 (owner: 10Joal) [22:11:38] 10Analytics, 10DBA, 10Operations, 10Patch-For-Review, 10User-Elukey: Decommission old dbstore hosts (db1046, db1047) - https://phabricator.wikimedia.org/T156844#3782789 (10Nuria) @Nettrom Right, is not only that dashboard but all the ones that are feed data via reportupdater that needed the new configura... [22:22:59] (03CR) 10EBernhardson: "overall this looks reasonable. A couple suggestions but not really required. You might also want to see the RFC giuseppe is working on: ht" (033 comments) [analytics/superset/deploy] - 10https://gerrit.wikimedia.org/r/392933 (https://phabricator.wikimedia.org/T166689) (owner: 10Ottomata) [22:33:18] (03CR) 10EBernhardson: Deploy repo for superset (python) (031 comment) [analytics/superset/deploy] - 10https://gerrit.wikimedia.org/r/392933 (https://phabricator.wikimedia.org/T166689) (owner: 10Ottomata) [23:15:38] (03CR) 10Nuria: Grow RestbaseMetrics spark job to count MW API (031 comment) [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/392700 (https://phabricator.wikimedia.org/T176785) (owner: 10Joal) [23:48:13] (03PS20) 10Mforns: [WIP] Add scala-spark core class and job to import data sets to Druid [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/386882 (https://phabricator.wikimedia.org/T166414) [23:52:36] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Add scala-spark core class and job to import data sets to Druid [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/386882 (https://phabricator.wikimedia.org/T166414) (owner: 10Mforns)