[00:41:36] Analytics, EventBus, Wikimedia-Stream: Public Event Streams - https://phabricator.wikimedia.org/T130651#2229653 (Krinkle) Below is a brain dump of some notes and ideas from earlier today. ---- Name ideas: wiomi, kiomi, kasoggi, kasiki **Wiomi** ([waɪˈoʊmɪ](https://en.wikipedia.org/wiki/Help:IPA_fo... [04:42:40] Analytics, Beta-Cluster-Infrastructure: deployment-fluorine puppet failure due to '/usr/sbin/usermod -u 10003 datasets' returned 4: usermod: UID '10003' already exists - https://phabricator.wikimedia.org/T117028#2229789 (Krenair) [08:10:37] gooooood morning [08:10:44] mforns: o/ [08:10:47] you there? [08:10:53] elukey, yes! [08:10:58] helloooo [08:10:58] morning :] [08:11:01] :) [08:11:07] so I have one thing to ask [08:11:11] if you have a bit of time [08:11:16] sure! [08:12:10] it seems that bromine (a ganeti VM) is a backend used by the Varnish Misc cache (where our websites are) to serve static stuff.. and I am checking if it is possible to move temporarily our websites in there for the stat1001 reimage [08:12:28] aha... [08:12:35] would you have time to write a very simple html sorry page for the analytics team? [08:12:59] I can do it but I am afraid that will look horrible [08:13:01] :D [08:13:09] sure can do :] [08:13:13] \o/ [08:13:19] at what time do you need it? [08:13:31] all right so let me check if this is possible first, then I'll give you more details [08:13:40] ok [08:13:43] I am planning to make the switch on Tuesday more or less [08:14:00] oh, ok [08:14:15] so yeah nothing super urgent [08:14:29] ok, I'll put it in my todo list [08:47:37] mforns: confirmed, it seems possible and easy to do.. but I'd need a gerrit repo with the page [08:47:41] to make everything smooth [08:47:53] elukey, aha [08:48:08] it's just an html file? [08:48:47] can you pass me the phab task you're working for me to have context please? [08:49:09] which sites are we planning to patch with the sorry page? [08:50:31] sure! all the ones on stat1001 [08:50:50] https://phabricator.wikimedia.org/T76348 [08:51:12] datasets.w.o, stats.w.o, metrics.w.o [09:21:53] * elukey just discovered that stat1001 has apache 2.2 installed... sigh [09:22:55] elukey: one of the many reasons for that jessie upgrade of yours :-) [09:23:38] !log installing ircbalance bugfix updates (preventing massive logspam on some systems) [09:23:39] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log, Master [09:24:37] moritzm: yesss! Just wondering if all the httpd directives are ok on the erb files.. [09:24:47] (Also I think that the log was meant to #ops?) [09:25:34] ah, right. thanks [09:31:39] Analytics, EventBus, Wikimedia-Stream: Public Event Streams - https://phabricator.wikimedia.org/T130651#2230037 (Afandian) Great to see thinking on this moving forward. Can I put in a vote for the most portable possible technology for connecting? For starters, there are current clients to RC_Stream i... [09:35:05] ah yesss apache erb files full of old directives [09:35:06] what a joy [09:35:58] elukey: Welcommmmmme to teh fantastic world of oldies wikimedia ! [09:36:21] I'm pretty sure the old 2.2 stuff is still around, after all we still need to run apache on precise for CI and a few other systems [09:41:07] moritzm: yeah I saw some of 2.2 around looking into cronspam (https://phabricator.wikimedia.org/T132661) [09:41:17] joal: thank youuu [09:42:36] joal: aqs is again throwing 500s, sigh [09:43:15] elukey: read rate * 2 [09:50:51] so include apache::mod::access_compat should do the trick, maybe we don't need new directives.. even if mod_authz_host should be used (Require ftw) [12:19:19] all right I have something vaguely working for stat1001 maintenance (https://gerrit.wikimedia.org/r/#/c/284878), will try to look for some review [12:22:55] I was wondering if bohrium (where we host piwik) would be fine as target host [12:26:03] Analytics-Kanban, Operations, Patch-For-Review: Upgrade stat1001 to Debian Jessie - https://phabricator.wikimedia.org/T76348#2230227 (elukey) Summary: 1) we decided to have a maintenance page hosted somewhere to avoid connection timeouts during the reimage. I created https://gerrit.wikimedia.org/r/2... [12:28:09] * elukey lunch! [13:53:13] mforns: https://gerrit.wikimedia.org/r/#/c/284878 kinda ready, let me know what you think whenever you have a min :) [15:05:28] elukey, I'm working in the maintenance page, should I include that in the puppet code? where? [15:06:22] mforns: you rock thanks :) [15:06:27] so I thought to use git::clone { 'analytics/websites_maintenance': in the code [15:06:39] so if you create the repo in gerrit it will be fine [15:06:58] I see [15:07:01] cool [15:07:49] elukey, I have not seen the code, but it would be cool that the http code returned by the server when returning the maintenance page is 503 [15:08:46] is that configurable from apache? [15:08:57] mforns: mmmmmmm could be an idea yeah [15:09:05] like setting an errordocument [15:13:21] elukey, how long is the outage going to be +/-? [15:17:51] ah if everything goes smoothly let's say 3/4 hours [15:20:12] ok [15:21:43] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2230580 (Yurik) @brion, the pretty printing is for diffs - makes them much easier to digest. A while ago... [15:21:47] do you like the idea of this maintenance page? [15:21:51] mforns: --^ [15:22:02] elukey, sure [15:22:43] it's the correct thing to do, no? [15:23:31] yeah theoretically it is, connect timeouts are horrible :( [15:25:55] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2230588 (brion) Re headers -- yeah need to distinguish between header labels (i18nable text) and column... [15:26:28] Analytics-Cluster, RESTBase-Cassandra: Evaluate TimeWindowCompactionStrategy - https://phabricator.wikimedia.org/T133395#2230591 (Eevans) [15:26:49] Analytics-Cluster, RESTBase-Cassandra: Evaluate TimeWindowCompactionStrategy - https://phabricator.wikimedia.org/T133395#2230605 (Eevans) p:Triage>Lowest [15:34:41] elukey, do you have rights to create a repo in gerrit? if so, please can you do it for me? [15:34:47] otherwise I can put the file as a gist? [15:35:10] mforns: never done it but I'll try [15:36:36] mforns: https://gerrit.wikimedia.org/r/#/admin/projects/analytics/websites_maintenance [15:36:46] thanks! [15:39:14] elukey, I think it's ready. it's rreeeally simple though. question: where should I put it in the repo? which path? I don't see that in the puppet change [15:43:35] ah right, I think that we can either go for README + docroot or just files directly [15:44:58] elukey, is this going to be reused for other outages? I hardcoded the task's URL... [15:45:50] possibly yes, but we can change it easily :) [15:46:46] mforns: i would not think much about reusage [15:47:00] mforns: can be changed as needed be for other outages [15:47:51] +1 [15:48:21] huh, did Prince's death overwhelm the pageview tool? 0 views for april 21 in https://tools.wmflabs.org/pageviews/#project=en.wikipedia.org&platform=all-access&agent=user&range=latest-20&pages=Prince_(musician) cc mforns MusikAnimal kaldari [15:49:07] ok [15:49:37] elukey, so just index.html? [15:49:56] nuria_: yeah [15:50:04] ahahha sorry [15:50:12] hehe [15:50:16] ok [15:50:16] my brain just didn't work correctly [15:50:18] HaeB: it is likely data is not loaded yet [15:50:47] Analytics, DNS, Operations, Traffic: Create analytics.wikimedia.org - https://phabricator.wikimedia.org/T132407#2230693 (Nuria) This is what I think needs to be done here to resolve this ticket as soon as possible: (cc-ing @BBlack and @Ottomata for confirmation) - host analytics.wikimedia.org... [15:51:08] nuria_: no, it already shows april 21 data for other pages, e.g https://tools.wmflabs.org/pageviews/#project=en.wikipedia.org&platform=all-access&agent=user&range=latest-20&pages=Mononymous_person [15:53:00] (PS1) Mforns: Add maintenance page for oncoming outage [analytics/websites_maintenance] - https://gerrit.wikimedia.org/r/284910 (https://phabricator.wikimedia.org/T76348) [15:54:26] elukey, ^ [15:54:30] (also, it's not that april 21 is missing in https://tools.wmflabs.org/pageviews/#project=en.wikipedia.org&platform=all-access&agent=user&range=latest-20&pages=Prince_(musician) - it shows 0 views for april 21) [15:55:40] HaeB: I see 5,808,147 for 21 April [15:55:47] (CR) Elukey: [C: 1] Add maintenance page for oncoming outage [analytics/websites_maintenance] - https://gerrit.wikimedia.org/r/284910 (https://phabricator.wikimedia.org/T76348) (owner: Mforns) [15:56:18] MusikAnimal: i see 0 ;) [15:56:22] the pageviews API takes about a day to populate, if the value is zero the tool should show nothing (and not zero), as as to indicate the data isn't available yet [15:56:25] i'm using chrome [15:56:57] HaeB: Windows/OSX? try opening it in incognito mode [15:57:06] linux [15:57:15] in incognito mode [15:57:21] should be fine nonetheless [15:57:30] does it actually say zero or is it blank for 21 April? [15:58:37] it says 0 [15:58:51] very odd [15:58:54] HaeB: pageview api is returning dat ajust fine: [15:58:58] https://www.irccloud.com/pastebin/7Fs0Dpi2/ [15:59:02] *data [15:59:05] "4/21/2106: 0" on hover [15:59:36] as MusikAnimal said there are 5816910 views as of yesterday [15:59:38] I assume you system time is correct? 22 April 16:00 GMT [15:59:44] *your [16:00:07] the JavaScript looks at your current time to determine whether to show zero or blank for the most recent day [16:00:10] nuria_: yes, i assume it's the tool (already checked pageview-hourly) [16:00:18] MusikAnimal: yes [16:00:19] but either way, if the data is there (which it is), it should show it [16:00:36] Analytics, Hovercards, Reading-Web-Sprint-71-m: Verify X-Analytics: preview=1 in stable - https://phabricator.wikimedia.org/T133067#2230710 (dr0ptp4kt) @nuria, awesome - it sounds like you'll have VCL that will classify preview=1 when client-initiated inside of X-Analytics. Do I understand correctly... [16:00:57] HaeB: do you know to check for JS errors? if not please see https://en.wikipedia.org/wiki/Wikipedia:Reporting_JavaScript_errors [16:01:03] maybe there's something breaking [16:02:35] nuria_: I see 8808147, not 5816910: https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/en.wikipedia/all-access/user/Prince_(musician)/daily/2016040200/2016042100 [16:02:37] MusikAnimal: no errors in the JS console [16:02:48] and it's the same under firefox [16:03:12] MusikAnimal: all-agents on mine: https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/en.wikipedia/all-access/all-agents/Prince_(musician)/daily/2016042000/2016042100 [16:03:25] oh okay, had mine set to User [16:04:09] HaeB: could you check the network log? the last entry is the API request. Right-click on "2016042100" and select "Copy link address" and paste here [16:05:22] Analytics-Kanban: Submit an issue and a PR to fix pageview.js bugs - https://phabricator.wikimedia.org/T133400#2230718 (mforns) [16:06:07] Analytics-Kanban, DNS, Operations, Traffic: Create analytics.wikimedia.org - https://phabricator.wikimedia.org/T132407#2230731 (Nuria) [16:06:58] Analytics-Kanban: Create maintenance page for stat1001 maintenance - https://phabricator.wikimedia.org/T133401#2230734 (mforns) [16:07:27] mforns: if you find a sec: https://gerrit.wikimedia.org/r/#/c/284834/1 [16:07:56] oh nm, madhuvishy reviewed it [16:07:57] nm! [16:09:02] https://www.irccloud.com/pastebin/eQ3JIWvh/ [16:09:21] MusikAnimal: ^ [16:09:54] looks like it did not query for april 21, but included it in the chart nevertheless [16:09:59] could you give me the URL though? [16:10:12] Analytics, Hovercards, Reading-Web-Sprint-71-m: Verify X-Analytics: preview=1 in stable - https://phabricator.wikimedia.org/T133067#2230756 (Nuria) >Do I understand correctly that would apply equally to the use case embodied by this task, which is where the browser does the header enrichment? Correct [16:10:20] https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/en.wikipedia/all-access/user/Prince_(musician)/daily/2016040200/2016042100 [16:10:24] Analytics, DBA: Set up auto-purging after 90 days {tick} - https://phabricator.wikimedia.org/T108850#1532166 (jcrespo) p:Triage>Normal [16:10:38] HaeB: I clearly see 5808147 as the last entry [16:10:51] from the URL, not from your pastebin [16:11:24] either way it should not say zero, it should be blank [16:11:46] that might actually be a bug, but the thing I'm most interested in is why I'm getting data back from the API that you are not getting [16:12:00] strange, for me 2016042000 is the last entry, even when opened separately in the browser [16:12:27] baffling [16:12:34] and you did try in incognito? [16:12:59] yes [16:13:37] and just tried via TOR (ie. from a different IP address) and there i indeed get 2016042100 as the last entry [16:14:16] ok need to work on something else right now, but can file a bug later today [16:14:36] ok [16:14:42] sorry! [16:15:22] for anyone interested in cross-language Wikipedia views for Prince (musician): http://tools.wmflabs.org/langviews/?project=en.wikipedia.org&platform=all-access&agent=user&start=2016-04-21&end=2016-04-21&sort=views&direction=1&page=Prince_(musician) [16:16:03] oh, i see it's not your fault probably [16:16:13] caching issues of the API? [16:16:32] perhaps [16:16:34] actually [16:16:48] again, the weird thing is that the tool gives me correct April 21 data for other pages [16:17:38] if you go back to the network log, it should say "(from cache)" under the "Size" column [16:18:20] but I think so long as you have the same options, the server caches that response for everyone [16:18:26] and I'm seeing the right data [16:19:30] for all Wikipedias: 7,608,953 pageviews of Prince_(musician) for 21 April [16:22:56] MusikAnimal: "size" column only says "736 B", no "from cache" [16:23:07] really weird [16:23:08] ...also, i meant server-side caching [16:23:21] (CR) Elukey: [C: 2] Add maintenance page for oncoming outage [analytics/websites_maintenance] - https://gerrit.wikimedia.org/r/284910 (https://phabricator.wikimedia.org/T76348) (owner: Mforns) [16:23:29] (CR) Elukey: [V: 2] Add maintenance page for oncoming outage [analytics/websites_maintenance] - https://gerrit.wikimedia.org/r/284910 (https://phabricator.wikimedia.org/T76348) (owner: Mforns) [16:23:35] right, but it would cache that for everyone, at least I would think [16:23:43] halfak: o/ [16:25:07] MusikAnimal: if it's a consolation, at http://tools.wmflabs.org/langviews/?project=en.wikipedia.org&platform=all-access&agent=user&start=2016-04-21&end=2016-04-21&sort=views&direction=1&page=Prince_(musician) i'm seeing the correct number too ;) [16:25:27] that's a different request [16:25:30] let's see... [16:25:41] elukey: do those vhosts just override the existing ones? [16:25:56] HaeB: try this one http://tools.wmflabs.org/pageviews/#project=en.wikipedia.org&platform=all-access&agent=user&start=2016-04-21&end=2016-04-21&pages=Prince_(musician) [16:26:41] or even try clicking on latest 10, latest 30, etc [16:26:48] MusikAnimal: that one is fine (5,767,168) [16:26:51] those will make different requests [16:26:53] awesome [16:27:33] well, or even more vexing [16:28:30] ottomata: I am testing them in labs.. the idea would be to install them on bohrium, where you have other ones [16:28:36] (piwik) [16:29:56] MusikAnimal: Just tried the multi-language tool: its great ! [16:30:06] thanks [16:30:18] it will be better once the Cassandra backend stuff is improved [16:30:31] right now I have to enforce throttling or else you'll get a lot of missing data [16:31:00] MusikAnimal: Yes, I understand [16:32:00] right, you were the one I was talking to about this! [16:32:01] sorry [16:32:13] MusikAnimal: np :) [16:33:50] I guess also sometimes Wikidata says a given article exists on some wiki and it doesn't, as you see with Price_(musician) [16:34:36] I thought those entires were added/corrected by bots [16:35:15] *entries [16:36:58] elukey: and change misc varnish to serve sites from there? [16:37:50] Analytics-Cluster, RESTBase-Cassandra: Standardized Cassandra dashboards - https://phabricator.wikimedia.org/T133403#2230822 (Eevans) [16:37:54] elukey: i wonder if we can do this more generically! [16:38:01] maybe we can make a more generic maintenance/downtime page [16:38:05] put it on some very generic host [16:38:28] Analytics-Cluster, RESTBase-Cassandra: Standardized Cassandra dashboards - https://phabricator.wikimedia.org/T133403#2230834 (Eevans) p:Triage>Low [16:38:29] and make it useable for any site downtime via puppet modification of varnish [16:39:06] ottomata: that's the idea :D [16:39:50] the only requirement is deploying the role statistics::web::maintenance on a health host behind cache::misc [16:40:31] and then flip the varnish director's settings to point to the new host [16:40:36] elukey: ja maybe we can make this more a misc::maintenance class [16:40:39] rather than make it stat related [16:40:41] ahhhh [16:40:44] do we even need special vhosts for this? [16:41:09] it could just be a generic maintenance catchall site...or it could be one vhost that is configured via puppet with multiple ServerAlias es [16:41:13] something like [16:41:25] Analytics: Implement Pages Created & Count of Edits full vertical slice - https://phabricator.wikimedia.org/T131779#2230836 (JAllemandou) After reviewing charts and talking with @ezachte, here is a schema-like view of data that would be needed to replicate most of the useful charts: Dimensions: - wiki... [16:41:37] misc::maintenance { sites => ['datasets.wikimedia.org', stats.wikimedia.org'] ... [16:41:49] (that's a bad class name, but ja) [16:42:20] I thought about serveralias but I didn't like it at first, but I don't really like the multiple vhosts either [16:42:24] mmmm [16:42:37] what about starting with statistics and then talk with brandon about a generic one? [16:45:32] Analytics, Collaboration-Team-Interested, DBA, Notifications: Purge all Schema:Echo data after 90 days - https://phabricator.wikimedia.org/T128623#2230846 (jcrespo) When is this taking effect (the 90-day deadline)? Even if data is deleted, I have to make sure tables are also deleted on all servers. [16:59:26] elukey: ok, let's do a single vhost, ServerAlias is fine. can we do a simple index.html page in puppet too, rather than making a repo? [17:01:29] ottomata: doing the single vhost atm, but I like the idea of the repo (also already done :P) [17:01:48] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2230912 (Yurik) I'm totally ok to bikeshed about the naming: * for ID, it will be a list of strings name... [17:03:43] ha, why repo? [17:05:50] ottomata: didn't want to mess with puppet, plus git::clone was super handy [17:09:10] hmm, elukey i guesssssssss, hm. [17:09:33] it is just a single file though, and you are already messing with puppet to deploy it, HMMMM Errrrrr [17:09:48] i don't really like it but not enough to block it so if you really like it i guess its ok! [17:11:15] elukey: when do you think you'll have time to backport the memcached package from sid? [17:12:22] ottomata, I reviewed your patch, I think it's fine, but maybe you want to add another line into the try block, so that the code is also rebust agains validation errors [17:12:29] ? [17:13:11] mforns: ? [17:13:16] ottomata, yes? [17:13:29] validation of the schema? [17:13:38] or validation of an event? [17:13:39] ottomata, validation of the scid [17:13:49] see the comments :] [17:14:03] oh ok...:) [17:15:10] ori: o/ later next week I should I have time, but I have never done it so I might be slower than necessary [17:15:22] elukey: cool, np, thanks! [17:15:46] ottomata: yep it is a single file but it might grow etc.. in the future, this is why I wanted to avoid putting stuff in puppet.. it is technically a website [17:15:54] ori: would it be fine for you? [17:15:58] I also updated the phab task [17:16:09] yep, totally fine [17:16:12] elukey: do you plan to add code to change varnish stuff in this patch too? [17:16:16] i'm wondering how that happens [17:16:34] elukey: the more I review this, the more I think it should be generic, you're going to have to touch varnish anyway [17:16:44] ah sorry i missed https://phabricator.wikimedia.org/T129963#2230527, will reply. thanks! [17:16:53] and, it would be much nicer to be able to parameterize the misc sites which get served maintenance pages in one place [17:16:54] ori: super thanks! [17:17:01] Analytics: Implement Pages Created & Count of Edits full vertical slice - https://phabricator.wikimedia.org/T131779#2230933 (JAllemandou) Sorry, messed up in the tasks, commented on the wrong one. [17:18:38] ottomata: nono I don't plan to add that part of the code, we'd still need to decide where to put the maintenance role first [17:18:48] but the change is one line in the varnish VCL [17:19:29] I'd like to make it generic but at the same time I want to make it working for us first, then bring the case for a more generic one for misc etc.. [17:19:44] otherwise I might end up creating a super big thing that we'll use only once [17:19:52] does it makes sense? [17:19:59] ja, but in doing so we've created a new repo, are cloning it onto a ~random host into /srv/analytics [17:20:06] if we chagne we'll have to clean all that up [17:20:14] that argument makes sense though [17:20:14] hm [17:20:33] but, you'll have to change it in the VCL, it seems pretty silly to have to change it in so many places [17:20:54] edit sites (serveraliases) to list the disabled sites, edit varnishvcl for those sites [17:20:55] hm [17:21:02] elukey: do you even need ServerAliases at all? [17:21:16] could you just do this on a non named vhost, ? [17:21:32] just use a custom port maybe? [17:21:43] then any domain that gets sent there from varnish would get the content [17:22:23] Analytics-Kanban: Examine wikistats reports, make a summary of the most granular data needed that would serve all reports - https://phabricator.wikimedia.org/T131783#2230955 (JAllemandou) After reviewing charts and talking with @ezachte, here is a schema-like view of data that would be needed to replicate mo... [17:22:32] ottomata: I think that it is better to specify the domains, otherwise we might get weird httpd behavior. Theoretically this might become a default vhost and it could serve the page for $everything [17:22:42] naw, make it on a custom port [17:22:53] custom port? [17:23:11] ja, not port 80 [17:23:47] don't even ahve to do it with apache if you don't want to [17:23:50] nginx would work too [17:24:01] but for apache (there is probbably puppet stuff to do this) [17:24:13] and what would be the end goal? [17:24:21] Listen 8081 [17:24:21] NameVirtualHost *:8081 [17:24:21] [17:24:21] ... [17:24:25] with no ServerName or ServerAlias [17:24:37] elukey: not having to hardcode or parameterize the sites [17:24:52] right now, say we add analytics.wikimemdia.org shortly [17:25:02] someone will have to remember to update this to that it works for that too [17:25:42] your goal is go make a maint page show up if someone visits those sites [17:25:53] and, that maint page needs to be hosted on a different server [17:25:58] so, you'll have to change the varnish routing [17:26:10] ottomata: I agree about the erb parametrization but not about the extra port, because it will need also ferm rules etc.. I mean, it is right to specify what are the domains to serve from apache IMHO [17:26:14] the site itself doesn't need any knowledge of what site is supposed to have maintenance [17:26:42] you are already adding a ferm rule [17:26:45] so that ain't no thing [17:26:59] why does apache need to know the domains? you aren't varying the content based on the domain [17:27:10] and even if you were, you could do that with .js or something [17:27:40] all you want is a simple maintenance page [17:27:48] heck ha, i'd probably even just do it at varnish level [17:27:53] apache doesn't need to, I said that IMHO it is a good practice, that's it :) [17:27:56] varnish is already configured to serve error pages [17:28:00] if it odesn't know about a domain incoming [17:28:04] those are in puppet (pretty sure) [17:28:15] why stand up a special server for this at all? [17:28:23] this makes sense [17:28:32] might be something to ask to Brandon [17:28:37] ottomata: traffic channel? [17:29:26] k! [17:29:56] Analytics: fr-tech request: browser version breakdown by country - https://phabricator.wikimedia.org/T133407#2230963 (DStrine) [17:33:38] (CR) Ottomata: [C: 1] "Can merge if you need me to! Just lemme know." [analytics/statsv] - https://gerrit.wikimedia.org/r/284836 (https://phabricator.wikimedia.org/T132770) (owner: Krinkle) [17:35:28] mforns: You will probably kill me I know, but varnish already serves a static error page for backend fetch failrues [17:35:34] I mean, misc ones [17:35:38] so we might just use it [17:35:54] without going through the whole mess [17:35:57] elukey, hehehehe no problem [17:36:11] it's better like this [17:36:19] mforns: feel really stupid sorry, I didn't think about the easiest choice :( [17:36:37] xDDDDDD no problem really, had a nice time doing it [17:37:57] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2230979 (Jdforrester-WMF) >>! In T120452#2230912, @Yurik wrote: > I'm totally ok to bikeshed about the n... [17:40:22] mforns: I spent 3/4 hours today on it [17:40:31] * elukey fires himself [17:44:06] elukey: don't be too hard on yourself :) [17:44:18] a-team, I'm off for diner ! [17:45:42] byeee [17:46:43] laers!~ [17:47:11] haha, man i've spent so many hours on things to realize there are better ways [17:47:18] sometimes you just can't know until you do it! [17:47:22] so is not time wasted :) [17:47:43] bye joal [17:48:18] +1 spending time on things :] [17:59:09] Analytics-Kanban, Operations, Patch-For-Review: Upgrade stat1001 to Debian Jessie - https://phabricator.wikimedia.org/T76348#2231122 (elukey) So after reviewing the change with @Ottomata we realized that it would be much easier to add a rule in VCL's cache::misc to enable/disable a backend with a fla... [18:01:05] a-team: logging off! Talk with you on tuesday! [18:17:11] Analytics, EventBus, Patch-For-Review: Continuous Integration for mediawiki/event-schemas - https://phabricator.wikimedia.org/T124438#2231187 (Legoktm) Open>Resolved a:Pchelolo https://integration.wikimedia.org/ci/job/npm-node-4.3/7094/console [18:17:43] Analytics, EventBus, Patch-For-Review: Continuous Integration for mediawiki/event-schemas - https://phabricator.wikimedia.org/T124438#2231194 (Pchelolo) Thank you @Legoktm [18:21:35] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2231244 (brion) @yurik: The W3C CSV on the Web working group's metadata model recommendation refers to "... [18:41:25] Analytics, EventBus, Wikimedia-Stream: Public Event Streams - https://phabricator.wikimedia.org/T130651#2231337 (Krinkle) The server will speak Socket.IO protocol, just like RCStream. So while the interface might be slightly different it is definitely not an increase in barrier. In fact, by upgrading... [18:42:28] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2231340 (Jdforrester-WMF) >>! In T120452#2231244, @brion wrote: > @Jdforrester-WMF data cube sounds awes... [19:13:20] Analytics, EventBus, Wikimedia-Stream: Public Event Streams - https://phabricator.wikimedia.org/T130651#2231404 (Afandian) Cool, thanks for considering my suggestion. It was based on the incorrect assumption that you would be looking to replace Socket.IO. [20:37:36] ottomata: IRC is EMPTY AGAIN [20:38:35] (CR) Nuria: "Nice. Thanks for doing this." [analytics/websites_maintenance] - https://gerrit.wikimedia.org/r/284910 (https://phabricator.wikimedia.org/T76348) (owner: Mforns) [20:42:53] nuria_, not empty any more :] [20:43:01] so lonelyyyy [20:44:07] nuria_, BTW, Neil responded to my peer review request saying that he will totally do that, but that really he has very few to say [20:44:45] nuria_, should I change? [20:48:16] hi nuria_ i am here! [20:56:13] Analytics: fr-tech request: browser version breakdown by country - https://phabricator.wikimedia.org/T133407#2230963 (Nuria) >Fr-tech is trying to understand how many wikipedia users are able to use local storage. We are interested in a browser version percentage breakdown by country. I think you will get p... [20:57:20] Analytics, Analytics-Cluster: Standardize use of refinery_path over oozie_path in all refinery oozie property files - https://phabricator.wikimedia.org/T133206#2231561 (Ottomata) [20:57:48] ottomata: hola amigo, I just saw all your eventlogging updates in my new vagrant [20:59:06] cool! uhhh oh for the role? [20:59:07] nice! [21:00:01] mforns: just merged your schema error thing [21:03:19] ottomata, thaaanks! [21:05:03] also, responded to your comments and put new patch up [21:05:10] both good points, i like! [21:05:11] thank you [21:05:31] :] will look at it now [21:05:54] ottomata: mw is still on port 8080 on vagrant internally right? [21:05:57] https://www.irccloud.com/pastebin/HxxeRgiC/ [21:06:38] it's the ":::8080" bind [21:07:12] things that bind to both IPv4 and IPv6 show as IPv6 in netstat [21:07:51] bd808: is it normal that my browser redirects gfrom http://localhost:8080 to http://dev.wiki.local.wmftest.net:8080/wiki/Main_Page [21:07:58] yeah [21:08:23] *.local.wmftest.net resolves to 127.0.0.1 [21:08:53] the pretty name was added as the preferred name a few weeks ago [21:08:56] ok, so internally and externally vagrant is on 8080? or internally is on 80? [21:09:13] bd808: thanks btw [21:09:23] inside the VM its on port 80 [21:09:46] Vagrant handles the :8080 -> :80 mapping [21:09:58] k [21:10:06] ok so that hasn't changed [21:10:43] bd808: thanks again [21:10:48] np [21:31:35] ottomata, merged your patch [21:32:22] cool danke [21:33:00] bye team, have a nice weekend! [21:33:24] ottomata: wowow, vagrant setup from scratch with varnish in an hour and a halfffff, ta-ta-channn #iamninja [21:35:10] nice! [21:35:19] :) [23:33:33] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2231771 (Yurik) More important semi-bikeshed questions: * How should we store licenses? Is there a lice... [23:38:31] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#1854102 (TheDJ) And how do you request deletion ? One idea is to use a multipart contenthandler. The Pa... [23:58:55] Analytics, Commons, Multimedia, Wikidata, and 3 others: Allow tabular datasets on Commons (or some similar central repository) (CSV, TSV, JSON, XML) - https://phabricator.wikimedia.org/T120452#2231791 (Yurik) Can the talk pages be used for deletion requests?