[00:49:45] (PS12) Nuria: Mobile apps oozie jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/181017 [01:10:43] Analytics-Cluster: Geo-coding UDF - https://phabricator.wikimedia.org/T77683#969576 (Nuria) >wouldn't it make more sense to have a field that stores the name and version of the DB we're using? At this time the code is not pluggable to any db and assumes maxmind return types. I think that noting what db versio... [01:14:16] MediaWiki-extensions-Popups, Analytics-EventLogging: Popups EventLogging on TOC click fails validation - https://phabricator.wikimedia.org/T86378#969577 (Prtksxna) a:Prtksxna [02:27:37] (CR) Nuria: "Also, another concern:" [analytics/refinery] - https://gerrit.wikimedia.org/r/184193 (owner: QChris) [03:25:56] (PS13) Nuria: Mobile apps oozie jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/181017 [05:33:40] https://gerrit.wikimedia.org/r/184288 - added job for analytics-limn-language-data [05:33:57] (oh. milimetic must be Zzzz :)) [06:19:14] Analytics-Dashiki: Failure to retrieve a metric json file should not break the UI - https://phabricator.wikimedia.org/T85233#969677 (kevinator) p:Triage>High [06:28:40] Analytics-Visualization: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#969681 (kevinator) [06:28:41] Analytics-Wikimetrics: Story: EEVSUser adds 'Pages created' metric - https://phabricator.wikimedia.org/T70353#969678 (kevinator) Open>Resolved a:kevinator grooming backlog. This was done in September. [06:58:02] Analytics-Visualization: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#969702 (kevinator) [06:58:04] Analytics-Wikimetrics: Story: EEVSUser adds 'Edits' metric - https://phabricator.wikimedia.org/T70352#969699 (kevinator) Open>Resolved a:kevinator grooming backlog. This was done in September. [06:58:24] Analytics-Visualization: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#733702 (kevinator) [06:58:25] Analytics-Visualization: Story: EEVSUser selects Target Site breakdown - https://phabricator.wikimedia.org/T70473#969703 (kevinator) Open>Resolved a:kevinator grooming backlog. This was done in December with the Pageview metric. [07:00:18] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#969707 (kevinator) [07:00:48] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#733702 (kevinator) [07:02:00] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#733702 (kevinator) [07:02:37] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#969711 (kevinator) [07:02:54] Analytics-Visualization: Story: EEVSUser loads dashboard with a default view - https://phabricator.wikimedia.org/T70140#969712 (kevinator) grooming backlog. This was done in September. [07:03:26] Analytics-Visualization: Story: EEVSUser loads dashboard with a default view - https://phabricator.wikimedia.org/T70140#969713 (kevinator) Open>Resolved a:kevinator [07:03:27] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#733702 (kevinator) [07:04:39] Analytics-Dashiki: Story: EEVSUser selects ALL wikis - https://phabricator.wikimedia.org/T70478#969716 (kevinator) [07:05:00] Analytics-Visualization: Epic: EEVS Release Candidate - https://phabricator.wikimedia.org/T70350#969718 (kevinator) Open>Resolved a:kevinator grooming backlog. This was done in September. [07:18:42] Analytics-Wikimetrics: Deploy newly register metric to production - https://phabricator.wikimedia.org/T70446#969736 (kevinator) Open>Resolved a:kevinator grooming backlog. This was done in September. [08:17:17] Quarry: Number of queries shown in profile is wrong - https://phabricator.wikimedia.org/T86512#969867 (Schnark) NEW [10:02:35] (PS1) Hashar: Jenkins job validation (DO NOT SUBMIT) [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184306 [10:02:38] (CR) jenkins-bot: [V: -1] Jenkins job validation (DO NOT SUBMIT) [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184306 (owner: Hashar) [10:02:53] kart_: the flake8 job fails :-D [10:02:59] (CR) KartikMistry: "recheck" [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 (owner: KartikMistry) [10:03:01] (CR) jenkins-bot: [V: -1] Add .gitreview file [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 (owner: KartikMistry) [10:03:18] ah there is no tox.ini file [10:04:10] blah [10:04:20] (CR) Hashar: [C: 1] "We can force merge this change. Then you will have to add a tox.ini file as described on https://www.mediawiki.org/wiki/Continuous_integra" [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 (owner: KartikMistry) [10:04:27] kart_: we can force merge that .gitreview file though [10:05:19] Sure [10:05:22] Thanks! [10:06:15] (CR) Hashar: [C: 2 V: 2] Add .gitreview file [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 (owner: KartikMistry) [10:06:20] done [10:06:35] kart_: just get a tox.ini file now and you will be fine ( https://www.mediawiki.org/wiki/Continuous_integration/Tutorials/Test_your_python ) [10:09:54] (Abandoned) Hashar: Jenkins job validation (DO NOT SUBMIT) [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184306 (owner: Hashar) [10:13:25] * kart_ blinks. We don't have anything like mobile (Python) yet :D [10:13:54] hashar: how to do simple jenkins enabling without tox/python stuff? [10:15:00] kart_: we used to have Jenkins jobs that would just run the flake8 command [10:15:07] but we are moving toward using well known entry points [10:15:16] so jenkins just invoke the entry point (for python tox ) [10:15:17] https://www.mediawiki.org/wiki/Continuous_integration/Test_entry_points [10:15:30] the PHP repos will be migrated to composer, so Jenkins would just 'composer install && composer test' [10:15:47] this way, developers can add whatever checks they might need [10:17:18] hashar: nice. Look like lots of stuffs happening :) [10:18:03] hashar: for now tox.ini is enough? [10:18:10] yup [10:18:13] cool. [10:18:28] kart_: and https://www.mediawiki.org/wiki/Continuous_integration/Tutorials/Test_your_python should explain it all [10:18:33] if not, I am willing to amend it! [10:20:14] doing :) [10:21:11] (PS1) KartikMistry: Add tox.ini for Jenkins Job [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184313 [10:21:17] hashar: ^ [10:21:44] jenkins love! [10:22:04] hehe [10:22:29] kart_: the deps with -r{toxinidir}/requirements.txt are not really needed [10:22:37] but it is not doing any harm anyway :] [10:23:06] (CR) Hashar: [C: 2] "The -r{toxinidir}/requirements.txt deps is not needed right now, but it is a good place holder for later on if you start adding dependenc" [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184313 (owner: KartikMistry) [10:23:12] (Merged) jenkins-bot: Add tox.ini for Jenkins Job [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184313 (owner: KartikMistry) [10:23:22] (CR) Hashar: "recheck" [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 (owner: KartikMistry) [10:23:32] kart_: \O/ [10:23:35] Thanks! [10:23:46] I'll fix requirements.txt later today. [10:25:32] (CR) KartikMistry: [C: 2] "As per Amir's +1 and discussion with Dan." [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 (owner: KartikMistry) [10:25:38] (Merged) jenkins-bot: Correct list of wikis [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 (owner: KartikMistry) [11:56:58] (PS1) KartikMistry: Better tabs name and subheading [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184328 [11:57:56] Any reason, http://language-reportcard.wmflabs.org/ is not updated after https://gerrit.wikimedia.org/r/183788 ? [11:58:02] (name of wikis) [12:03:24] (CR) QChris: "> When trying to use the refined dataset I get:" [analytics/refinery] - https://gerrit.wikimedia.org/r/184193 (owner: QChris) [12:06:19] kart_: nbwiki does not exist. [12:06:31] Not sure but, when I try to go to [12:06:36] nb.wikipedia.org [12:06:42] it forwards to no.wikipedia.org [12:06:52] So maybe you meant nowiki instead of nbwiki? [12:07:05] (nowiki database exists) [12:07:44] Thanks. Fixing! [12:08:27] yw [12:12:20] (PS1) KartikMistry: Fix language code [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184331 [12:12:43] qchris: thing is that ISO code is nb :) [12:13:07] But fixed as of now. [12:13:53] Well .. what can we do if ISO does not follow WMF's wiki naming? :-P [12:15:23] But ... on second thought ... if you're looking for NB in ISO ... maybe you're not looking for nowiki. [12:15:30] Which Wiki are you looking for? [12:15:58] https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2#Decoding_table [12:16:09] lists "NB" as unassigned [12:16:14] kart ^ [12:17:04] kart_ ^ [12:17:16] https://en.wikipedia.org/wiki/Norwegian_language [12:17:56] Bokmal [12:19:19] ok. Then nowiki seems to be what you're looking for. [12:19:21] Cool. [12:19:26] https://en.wikipedia.org/wiki/Norwegian_Wikipedia [12:27:16] (CR) KartikMistry: [C: 2] "nb->no" [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184331 (owner: KartikMistry) [12:27:21] (Merged) jenkins-bot: Fix language code [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/184331 (owner: KartikMistry) [13:54:59] morning milimetric [13:55:31] morning Ironholds [13:58:52] Analytics-Engineering: Mobile product managers should have reports on session-related metrics from the Wikipedia Apps - https://phabricator.wikimedia.org/T86535#970748 (Ironholds) NEW [14:09:18] morning milimetric [14:09:34] hey kart_ I'm in a meeting for a couple hours [14:10:06] milimetric: np. Just 'greeting' :) [14:10:29] When you've time, how to add human readable title instead of 'content_translation_beta' in dashboard. [14:10:33] Q^ [14:52:16] Analytics-Cluster: Geo-coding UDF - https://phabricator.wikimedia.org/T77683#970854 (Ottomata) DarTar, I do like that idea though. When I incorporate this UDF into the refined table, I'll see if I can tag the line with the version of the db as well. [14:54:21] Good morning Andrew [14:55:11] morning ottomata :) [14:55:23] sucheta, do you have any need for books at all-staff? I can has a ton of Manguel. [14:56:48] morning ananthrk [14:56:50] Ironholds: [14:56:52] :) [15:31:44] ottomata1: Wanna trap chat real quick about the changes I uploaded? [15:32:18] gimme a few, then yes [15:32:38] k [15:33:14] brb [15:35:05] back [15:38:52] ok heading to trap [15:44:51] Analytics-Cluster, Analytics-Engineering: Mobile product managers should have reports on session-related metrics from the Wikipedia Apps - https://phabricator.wikimedia.org/T86535#970913 (kevinator) p:Triage>Normal [15:46:50] Analytics-Visualization, Analytics-Engineering: generate.py produced broken CSV files - https://phabricator.wikimedia.org/T86059#970917 (kevinator) [15:54:37] Analytics-Visualization, Analytics-Engineering: Analysts visualize Pageview 0.4 cube in Pentaho - https://phabricator.wikimedia.org/T86540#970928 (kevinator) NEW [15:58:30] (PS3) Gilles: Query image performance by upload time [analytics/multimedia] - https://gerrit.wikimedia.org/r/180136 (https://phabricator.wikimedia.org/T76035) [15:58:33] (CR) jenkins-bot: [V: -1] Query image performance by upload time [analytics/multimedia] - https://gerrit.wikimedia.org/r/180136 (https://phabricator.wikimedia.org/T76035) (owner: Gilles) [16:04:26] (PS4) Gilles: Query image performance by upload time [analytics/multimedia] - https://gerrit.wikimedia.org/r/180136 (https://phabricator.wikimedia.org/T76035) [16:06:01] (PS14) Nuria: Mobile apps oozie jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/181017 [16:08:05] ottomata1: should the refined datasets have a "SUCCESS" flag to indicate they are done processing too? [16:09:02] not necessarily, if we think they should we can add it, but since they they are completely controlled by hive, they won't appear until they are ready for querying in hive [16:13:56] Analytics-EventLogging: Add sanitized User-Agent to default fields logged by EventLogging - https://phabricator.wikimedia.org/T54295#970982 (ggellerman) [16:14:47] ottomata1: ok, i see. [16:15:25] Analytics-EventLogging, Analytics-Engineering: Validate JsonSchemaContent using MediaWIki core's handling - https://phabricator.wikimedia.org/T76432#970987 (kevinator) p:Triage>Normal [16:16:03] Analytics-EventLogging, Analytics-Engineering: Validate JsonSchemaContent using MediaWIki core's handling - https://phabricator.wikimedia.org/T76432#970989 (ggellerman) [16:20:09] ottomata: in my case I am using the datasource plain w/o adding anything related to success but i still get that the file is needed [16:20:17] https://www.irccloud.com/pastebin/R1B26w7N [16:21:10] Analytics-EventLogging: VE instrumentation is not showing up in databases - https://phabricator.wikimedia.org/T74173#971012 (Milimetric) Open>Resolved a:Milimetric Edit team has a dashboard that is fed data from the DB, so everything seems ok from our point of view: http://edit-reportcard.wmflabs.org/ [16:22:12] ottomata: job is 0023374-141210154539499-oozie-oozi-C [16:22:45] Analytics-EventLogging, Analytics-Engineering: EventLogging calling deprecated SyntaxHighlight_GeSHi::buildHeadItem - https://phabricator.wikimedia.org/T71328#971028 (kevinator) p:Triage>High [16:25:16] hm [16:26:44] hm, nuria, i think i need to set [16:26:46] on the dataset [16:26:58] [16:27:00] ootomata: want me to try? [16:27:02] yes please [16:27:06] k [16:27:18] i guess it is default to have _SUCCESS [16:28:39] Analytics-EventLogging: Two tests classes on testing efSchemaValidate - https://phabricator.wikimedia.org/T67546#971040 (ggellerman) [16:29:41] Analytics-EventLogging, Analytics-Engineering: translate.googleusercontent.com in webHost for some client-side events - https://phabricator.wikimedia.org/T67508#971042 (kevinator) [16:29:59] Analytics-EventLogging, Analytics-Engineering: translate.googleusercontent.com in webHost for some client-side events - https://phabricator.wikimedia.org/T67508#971046 (kevinator) p:Triage>High [16:31:47] (CR) Ottomata: [C: 2 V: 2] Adjust comment for Camus' whitelist to match the current setting [analytics/refinery] - https://gerrit.wikimedia.org/r/184190 (owner: QChris) [16:32:20] (PS2) Ottomata: Feed the refined text table from the unrefined text (not upload) table [analytics/refinery] - https://gerrit.wikimedia.org/r/184193 (owner: QChris) [16:32:28] (CR) Ottomata: [C: 2 V: 2] "Whoopsie!" [analytics/refinery] - https://gerrit.wikimedia.org/r/184193 (owner: QChris) [16:35:09] Analytics-EventLogging: Epic: ProductManager visualizes EL data - https://phabricator.wikimedia.org/T75068#971053 (kevinator) p:Triage>Unbreak! [16:35:40] Analytics-EventLogging: Epic: ProductManager visualizes EL data - https://phabricator.wikimedia.org/T75068#762969 (kevinator) p:Unbreak!>Normal [16:37:28] Analytics-Visualization, Analytics-Engineering: Analysts visualize Pageview 0.4 cube in Pentaho - https://phabricator.wikimedia.org/T86540#971062 (kevinator) a:mforns [16:38:27] Analytics-Visualization, Analytics-Engineering: Analysts visualize Pageview 0.4 cube in Pentaho - https://phabricator.wikimedia.org/T86540#971066 (ggellerman) [16:38:38] Ironholds, hi, I'm going to crate the cube for the new data, can you tell me where it lies? [16:39:42] mforns, sure, I just need to compress it to month-level [16:39:44] gimme 10m [16:39:55] Ironholds, of course :] [16:45:22] haw, nuria, sorry, the .hql file can (and does in other jobs) have a verb in it, just not the directories. generate_daily_uniques.hql was a fine name, and is consistent with other names we already ahve [16:45:40] ottomata: understood [16:49:37] ottomata: other problem is that "hours" are being passed by oozie like "00" [16:50:11] ottomata: so "hour=00" is the expected path instead of "hour=0" [16:50:16] ottomata: makes sense? [16:50:29] mforns, staging.pentaho04 [16:50:57] Ironholds, ok thanks! [16:51:53] np! [16:52:16] that's fine, no? oh. [16:52:17] that is aproblem? [16:52:18] hm [16:52:21] oh i see [16:52:53] ottomata: it is if we want to be consisten with oozie format [16:52:53] ah because the dataset doesn't match [16:52:54] hm [16:52:57] right [16:53:11] https://www.irccloud.com/pastebin/qCqnTKoW [16:53:18] note the "00" [16:53:55] ok yeha, [16:54:07] hm [16:54:26] since the dataset is explicitly within coorindator context, i think we can use coord:formatTime in the dataset definition to fix [16:55:56] Analytics-EventLogging: Unit tests for Event Logging - https://phabricator.wikimedia.org/T86543#971104 (Aklapper) [Please associate a project when creating a task - setting EventLogging here] @ggellerman: Also, if I was a volunteer, I'd have no pointers (documentation pages how to write unit tests) where to... [16:56:01] Analytics-EventLogging: Unit tests for Event Logging - https://phabricator.wikimedia.org/T86543#971106 (Aklapper) p:Triage>Volunteer? [16:56:01] nuria: since you are testing, can you try changing the ${HOUR} part of the path in datasets_refined.xml to [16:56:07] ${coord:formatTime(${HOUR}, "H")} [16:56:07] ? [16:56:08] maybe? [16:56:16] i guess we'd have to do this for the day and month too [16:56:25] ottomata: sure, will try and report [17:01:27] i've gotta run an errand during lunch [17:01:37] bbl [17:10:21] (CR) Ottomata: "- generate_daily_uniques.hql was a good filename for your hive query file." (4 comments) [analytics/refinery] - https://gerrit.wikimedia.org/r/181017 (owner: Nuria) [17:20:54] (CR) Gergő Tisza: [C: 2] Query image performance by upload time [analytics/multimedia] - https://gerrit.wikimedia.org/r/180136 (https://phabricator.wikimedia.org/T76035) (owner: Gilles) [17:21:04] (Merged) jenkins-bot: Query image performance by upload time [analytics/multimedia] - https://gerrit.wikimedia.org/r/180136 (https://phabricator.wikimedia.org/T76035) (owner: Gilles) [17:27:53] qchris: ytr? [17:29:18] ottomata: not having much luck with hour=${coord:formatTime(${HOUR}, "hh")}: or hour=${coord:formatTime(coord:nominalTime(), "hh" [17:31:23] ottomata: i do not think formating is supported by uri-template [17:32:02] nuria: what is the problem, maybe it is not [17:32:05] btw, don't you want "H" [17:32:06] not "hh"? [17:33:50] ottomata: ah wait, the coordinator and workflow do it differently [17:34:30] hm, yes? [17:34:37] not sure what you mean but yes? [17:37:15] ottomata: still i do not think formating is supported on the uri field: [17:37:19] https://www.irccloud.com/pastebin/cit384P3 [17:37:43] aye, what's the error? [17:37:54] Error: E1004 : E1004: Expression language evaluation error, Unable to evaluate [17:38:24] ottomata: let me do some digging to see if you can add formats there [17:41:33] ah, nuria [17:41:47] hmmm [17:41:49] i think you can't, hmm [17:41:50] but [17:42:01] ahaham jedi master [17:42:44] hmm, MMM [17:43:00] hmm, naw hm [17:43:03] don't know. [17:43:13] we might need to change the partition scheme [17:43:16] back to something that works with oozie :/ [17:43:50] crapo i think we do [17:44:07] but didn't qchris said he had used the dataset? [17:44:43] from oozie? [17:44:53] hm [17:44:58] ottomata: let's consult with qchris when he's back [17:45:10] yes, but maybe he didn't check that the single digit timestamps ran [17:45:16] ahhh [17:45:29] ottomata: do you know what job? [17:45:31] nuria: he had a coordinator running from his home dir on stat1002, right? [17:45:35] i was about to ask you that? [17:45:35] :) [17:47:40] ja dunno [17:47:40] hm [17:48:02] ottomata: ok, let me look at jobs running [17:48:33] i'm looking too [17:48:38] don't see anythingn owned by him that has it [17:52:00] ottomata: ok, let's pause on this for a sec then until qchris comes back cause otherwise if we change the partition scheme we might break the jobs he's running right? [17:52:58] yes, i don' thikn he's running any [17:53:24] but i'm ok with waiting if you are, he said he'd be back later this afternoon. i probably will be offline then, but the two of you can probably figure i tout [17:54:41] ok, let's wait on this until he's back, will try to catch him and I can move to help mforns with his vagrant stuff [17:57:30] ori: question about eventlogging devserver [17:57:59] ori: we always initialized it inside the vagrant db right? [17:58:28] ori: so it was publishing its port so it would be visible outside teh db itself [17:59:46] ? [18:01:08] ori: I was wondering if the eventloggingdevserver in vagrant [18:01:31] running in say port 8100 [18:01:56] was always visible outside vagrant or rather we used to run the devserver "outside" vagrant [18:02:16] ori: does question make sense? [18:02:45] outside vagrant [18:02:53] if you ask me [18:03:02] the logging url should also use port 8080 [18:03:11] and there should be apache config for routing it to the dev server [18:03:14] via mod_proxy [18:04:38] cc : mforns please see ori's comment above [18:04:51] hi! right [18:05:43] ori: why do we need the apache routing just to test in vagrant? [18:06:05] because people typically test their logging setup in their normal browser [18:06:13] and it's the browser that makes the eventlogging requests [18:06:19] so it has to be able to ping the devserver [18:06:43] we could forward a dedicated port for the devserver, btu that seems silly and wasteful [18:07:29] ori: but then ....the devserver is running outside vagrant but the apache config that routes ports is inside vagrant puppet? [18:08:05] it should run in vagrant [18:08:16] we only require that the devserver run outside vagrant because of this limitation [18:08:26] people would be happy to not have to also install eventlogging on their host env [18:12:37] ah ok, ori, get it now, [18:13:42] ori: so if we do the apache config the devserver should run inside vagrant correct? cc mforns [18:14:08] following [18:14:18] yes [18:15:53] ok nuria, I can do that. First, however I'm importing the new cube into Pentaho [18:16:08] kevin asked me to do so first [18:17:27] mforns:if it is for reserachers , they can do that too, right? [18:18:24] yes, but it seems this is priority, to be delivered this SF-morning [18:18:42] anyway nuria I'll finish in 15 minutes [18:24:29] Ironholds, the data is in Pentaho, could you have a look to see if it is OK? [18:25:45] nuria, I'm stopping for a quick lunch, and then will start with vagrant changes [18:25:53] mforns: sounds good [18:26:25] mforns, in meeting :( [18:26:38] ok Ironholds no rush :] [19:17:34] Ironholds: [19:17:38] you around? [19:20:11] tnegrin, yup [19:20:45] hey -- so on the weird rise in mobile/drop in page views [19:21:04] I was thinking we could use your newly minted dataset to see if we see this there also [19:21:36] Sure, although the link there isn't dependent on the pageviews count [19:21:43] Like, I've done session analysis to show it, too [19:22:15] (mobile sessions per user * events per mobile session) < (desktop sessions per user * events per desktop session) [19:23:12] ok -- I'm just thinking about the strange and noticeable shift in the last couple of weeks [19:23:47] agreed that the overall drop in PVs lines up [19:23:52] oh, I was pretty sure that was just christmas [19:24:07] people don't use desktop devices on holidays, they do use their phones [19:24:21] but we can dig into it, yeah [19:24:30] heh -- that's a good theory didn't think about that [19:24:54] tnegrin: I have a seen a similar pattern happening on summer vacation with mobile requests [19:24:59] cc Ironholds [19:25:14] we need to verify it but i agree with Ironholds [19:26:07] ok -- I'm sold -- this makes sense [19:26:21] it's a window into the future for sure [19:26:27] yup [19:26:32] and it has implications for fundraising particularly [19:26:44] huge [19:26:45] given when our season is [19:27:08] I think I'd like Ellery to give a talk about what he saw in FR this year [19:27:26] anyway -- thanks all for the insights [19:27:36] yeah [19:28:13] we see it on Dutch wiki too -- almost the same pattern [19:28:19] nlwiki [19:31:02] David Cameron's response to "Birmingham is entirely Muslim" - �When I heard this, frankly, I choked on my porridge and I thought it must be April Fools� day, This guy�s clearly a complete idiot.� [19:31:12] For once, I agree with my all-powerful leader. [19:32:52] kevinator: please note that labs is on better health and that you can now see metrics for rollingactiveeditor for enwiki: https://metrics.wmflabs.org/static/public/dash/#projects=ruwiki,itwiki,dewiki,frwiki,enwiki,eswiki,jawiki/metrics=RollingActiveEditor [19:34:23] nuria: wow, that’s awesome :-) [19:35:07] kevinator: have it in mind for the annotations, after springle fixes on labs looks like things are much better [19:37:20] (PS1) Milimetric: Update for January meeting [analytics/reportcard/data] - https://gerrit.wikimedia.org/r/184410 [19:37:25] nuria: are dewiki itwiki going to get better too? [19:37:36] (CR) Milimetric: [C: 2 V: 2] Update for January meeting [analytics/reportcard/data] - https://gerrit.wikimedia.org/r/184410 (owner: Milimetric) [19:38:07] kevinator: dewiki and itwiki data is there right? it was not missing [19:38:45] there was no data for those 2 over the weekend [19:38:47] kevinator: ah no, i see at the end they are not being calculated, we will need to look into why that is [19:39:11] kevinator: seems a problem only for itwiki though [20:17:38] mforns, I borked up the data construction [20:17:42] soo..we need to reimport ;p [20:18:07] Ironholds, np [20:18:23] * Ironholds goes to rebuild data [20:18:28] Ironholds, from what point, from the initial table? [20:19:07] hmn? [20:19:11] oh, I just need to smush the fields [20:19:26] (technical term OBVIOUSLY >.>) [20:19:31] xD [20:20:38] Ironholds, do you want to do the importing with me, in a hangout, so next time you do not depend on eng? [20:21:01] totally! Just give me 30s :) [20:21:07] and then we can step through it [20:21:34] Ironholds, OK [20:28:37] mforns, around [20:28:51] ok [20:29:04] Ironholds, do you want to do a hangout? batcave? [20:29:45] hangout wfm; just send me the URL! [20:30:11] Ironholds, https://plus.google.com/hangouts/_/wikimedia.org/a-batcave [20:30:54] mforns, wait one [20:30:55] okay, the numbers are actually just wrong [20:31:00] I need to work out why and regenerate :/ [20:31:02] tnegrin, ^ [20:31:09] oh [20:31:14] ok [20:31:44] womp womp [20:31:59] kudos to ellery for catching it, kudos reduced for ellery for never being on IRC and forcing the use of lossy, async communications methods to debug ;p [20:32:25] (PS2) Nuria: Prevent crashing ui when metric data is missing [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) (owner: Mforns) [20:32:41] Ironholds, let me know when I can help, I'll be around for 2 more hours :] [20:32:54] well, probably tomorrow ;p. I've made it whip-fast [20:33:03] by which I mean, it only takes 24 hours to go through 600 days of logs. [20:33:30] Analytics-Engineering: EL-Kafka pipeline that was disabled 6 months ago - https://phabricator.wikimedia.org/T86576#971775 (ggellerman) NEW [20:34:11] mforns: i know you had lost hope but i have CR-ed https://gerrit.wikimedia.org/r/#/c/182946, see if comments make sense [20:34:37] Ironholds: can you run it for a month and make sure the data is correct before the 24 hour run? [20:34:41] I was looking at it, now I understood what you were saying about the this and that [20:35:39] tnegrin, yep [20:35:43] I mean, I'll probably just run it for a day [20:35:47] I know what the day counts should look like [20:36:03] mforns: right, that wax needed before (in the *ahem* old times) when bind did not exist as a native method of function object [20:36:30] mforns: if you are good with that let's just merge that change [20:37:04] *that was needed [20:37:37] mforns: let me see that tests run, shouldn't hurt any but it does not hurt to verify [20:38:30] nuria, yes, I'm completely ok with how it looks now [20:39:20] (CR) Nuria: Prevent crashing ui when metric data is missing (1 comment) [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) (owner: Mforns) [20:39:27] Ironholds, OK let me know tomorow [20:39:36] sure [20:40:42] Ironholds: do you think you could follow this instructions and try to import the cube: https://wikitech.wikimedia.org/wiki/Analytics/Pentaho [20:40:57] Ironholds: that ay we make the rocess self-service [20:40:57] *process [20:41:19] do I have access to dan-pentaho.eqiad.wmflabs? [20:42:09] Ironholds: Try ssh-ing, if it doesn't work you can ask in wikimedia-labs to get access as you should have access to all analytics instances [20:42:34] sure [20:42:34] I'll do all of that, tomorrow [20:42:43] right now the code has a bug and the longer it takes me to isolate and fix it the longer before the run can kick off [20:42:49] mforns: actually even with patch #1 11 tests fail [20:42:59] Ironholds: THANKS! [20:43:01] so, while I appreciate everyone is trying to help, right now the most helpful thing is either not poking the bear or emailing the bear :D [20:43:15] nuria, wow [20:43:21] (evidently I'm a bear now. Maybe the Winnie the Poo kind? I do like honey.) [20:43:37] anyway. nuria, thank you for the doc pointers - bookmarked and will look into :) [20:43:40] Ironholds: I am not the best with metaphors but I think i understand [20:44:52] (CR) Nuria: "Just run tests but several are failing. We need to fix those before merging." [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) (owner: Mforns) [20:45:44] nuria, test pass for me [20:47:41] mforns: for me there are also failures in master [20:47:59] mforns: can you try master? [20:48:05] nuria, yes, trying [20:48:33] nuria, all tests pass in master for me [20:49:30] mforns: ok, must be me, let me retry [20:50:59] milimetric: can you try dashiki's tests and see if they pass? [20:51:10] nuria, maybe you need 'bower install'? or 'npm install'? [20:51:11] sure, one sec [20:51:48] nuria: yep, 33/33 pass, bower install didn't do anything (so I was already updated) [20:51:52] you having trouble? [20:52:32] milimetric, mforns ok solved now [20:52:35] i think i had a bad file [20:52:40] so they pass [20:53:03] fine [20:53:13] can I merge then? [20:53:28] (PS3) Mforns: Prevent crashing ui when metric data is missing [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) [20:53:42] (CR) Nuria: "Taking my last comment back, issue with my dev env." [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) (owner: Mforns) [20:53:47] mforns: we can merge that change then [20:53:51] ok [20:54:05] mforns: yes, one less on teh list [20:54:06] *the [20:54:18] (CR) Mforns: [C: 2 V: 2] Prevent crashing ui when metric data is missing [analytics/dashiki] - https://gerrit.wikimedia.org/r/182946 (https://phabricator.wikimedia.org/T85233) (owner: Mforns) [20:54:45] yea! thanks [20:55:06] nuria / mforns: have you ever run the qunit tests for event logging? Do those pass for you? [20:55:16] I've been staring at this too long, I think I need help [20:55:24]