[00:51:32] Analytics-Wikimetrics: Epic: WikimetricsUser shares a cohort - https://phabricator.wikimedia.org/T69461#1130857 (Fhocutt) [01:45:13] springle: yt? [01:47:08] nuria: hi! [01:47:18] springle: have time for 1 question? [01:47:39] Analytics-Tech-community-metrics, MediaWiki-Developer-Summit-2015, ECT-March-2015: Achievements, lessons learned, and data related with the MediaWiki Developer Summit 2015 - https://phabricator.wikimedia.org/T87514#1130942 (Rfarrand) More updates added, not quite finished. Comments so far? [01:48:43] nuria: yep [01:48:50] springle: k, this table: [01:48:55] https://www.irccloud.com/pastebin/RfMEwUHQ [01:49:10] has an id field w/o an auto_increment [01:49:58] is that something we removed? [01:50:11] nuria: i guess so? the master is the same [01:51:27] springle: ok, i do not understand how that can be, will research, as old EL code creates tables with id field with autoincrement and newer EL code (deployed this week) will create tables w/o that id [01:51:48] nuria: the only person i know of manually touching EL tables recently on the master is ori. perhaps he touched it. if not, some bug [01:52:09] i di reall something about him truncating a table and recreating, but idk the details [01:53:29] nuria: should we fix it, or do you wish to investigate first? [01:54:31] springle: I think i will investigate and let you know. [01:55:52] nuria: https://phabricator.wikimedia.org/T91031 might be a logical place to discuss this (as an aside). and if other tables may be affected [01:56:43] springle: ok, got it [01:58:40] springle: that is odd cause seems that it is not related to code [01:58:58] ori: yt? [09:08:32] Analytics, Analytics-Kanban: Turn off WP Zero's Limn-Dashboards & put up a "moved sign" - https://phabricator.wikimedia.org/T92920#1131227 (QChris) >>! In T92920#1130547, @kevinator wrote: > Spoke to Dan and this isn't as simple as dropping in an index.html where the current dashboards are. > > Easiest s... [09:47:56] Analytics-Tech-community-metrics, MediaWiki-Developer-Summit-2015, ECT-March-2015: Achievements, lessons learned, and data related with the MediaWiki Developer Summit 2015 - https://phabricator.wikimedia.org/T87514#1131281 (Qgil) Thank you, this is very interesting. I think the level of detail is eno... [10:32:45] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131462 (yuvipanda) NEW [10:33:14] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131469 (yuvipanda) I am temporarily moving the biggest of the files (51G eventlogging_processor-client-side-events.log.1) on to /srv. Someone who knows m... [10:59:32] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131505 (yuvipanda) The biggest logs seem to be filled with validation errors about https://meta.wikimedia.org/wiki/Schema:Edit [11:09:41] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131534 (yuvipanda) p:Triage>Normal [11:58:23] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131591 (yuvipanda) p:Normal>Unbreak! Free space is being gobbled up really fast still, and won't last more than a few hours. @nuria Is the eventl... [11:58:59] qchris: around? [12:00:18] ah, he isn’t... [12:00:19] I guess? [13:16:44] milimetric: hey [13:16:53] hi YuviPanda [13:16:54] milimetric: https://phabricator.wikimedia.org/T93185 [13:17:00] milimetric: might be causing EventLogging issues [13:17:07] (full disk) [13:17:09] oh coitus [13:17:12] yeah, just read it [13:17:14] hahaha [13:17:15] that's a big problem [13:17:26] yeah [13:17:51] * YuviPanda is on spotty internet atm [13:18:04] especially so because we just realized we might need to go manually repair all the old data [13:18:09] so throwing away the logs can't happen [13:18:27] (all the old data == the bad Edit schema data that we need for VE numbers) [13:18:49] thanks YuviPanda, I'll go have a look [13:19:00] milimetric: yup, I moved some logs to /srv [13:19:06] figured deleting the logs would be terrible [13:19:29] milimetric: yw. [13:19:52] Analytics, Analytics-EventLogging, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131723 (Milimetric) Thanks @YuviPanda for the emergency fix, I'm tagging our team and making this high importance. [13:20:08] Analytics, Analytics-EventLogging, Analytics-Kanban, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131724 (Milimetric) [13:25:11] Ironholds, hi! [13:26:22] Analytics, Analytics-EventLogging, Analytics-Kanban, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131736 (yuvipanda) Cool :) I also copied logs_02_06_onward folder onto /srv to make space as well. [13:26:33] Analytics, Analytics-EventLogging, Analytics-Kanban, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131737 (yuvipanda) Cool :) I also copied logs_02_06_onward folder onto /srv to make space as well. [13:27:15] milimetric: as ops on ‘ops clinic duty’, I have handed this off to you now :) [13:28:31] YuviPanda: acknowledged. Though I will state for the record that giving me ops things is like giving toddlers chainsaws. [13:29:04] milimetric: :D [13:29:43] hey mforns [13:29:58] milimetric, aw, no. Giving ME ops things is like giving toddlers chainsaws [13:30:05] giving you ops things is like giving a moody preteen a chainsaw [13:30:11] don't be so hard on yourself ;p [13:30:59] Ironholds, the isAppPageRequest method you linked me to, is not in the refinery repo? [13:31:12] erm [13:31:15] yes it is? [13:31:22] I'm looking at it literally right now ;p [13:31:32] it is, however, currently a subsection of isPageview and not exposed as its own UDF [13:31:43] Ironholds, in the compiled code in stat1002, I can find isPageview, for example, but not isAppPageRequest [13:31:49] Ironholds: pfft, you haven’t had leslie carr email you stern emails, have you? :P me and milimetric have :) [13:31:51] which is why I said I'd work on it this morning, because pulling it out means (pulling it out + UDF + unit tests) [13:32:12] mforns, yes, because it's a private method within isPageview [13:32:22] frankly I'd be pretty worried if you COULD find it ;p [13:32:23] Oj [13:32:25] oh! [13:32:32] it'd mean I screwed up [13:32:50] so, I'm gonna spend the morning pulling it out and writing unit tests and that'll be bundled into my big architectural change to refinery-core [13:32:55] and then it'll be available! \o/ [13:33:07] Ironholds, my bad, thx [13:33:17] (I'd make it its own distinct thing, but it'll be dependent on said big architectural change and so unless that gets CRd quick, they're going in the same patch [13:33:18] NP! [13:33:32] my bad too :D. [13:33:38] YuviPanda, actually, yes I have ;p [13:33:52] also, she's still deeply amused that her stern emails are talked of in such tones [13:34:19] Ironholds: publicly? :P [13:34:27] indeed! [13:35:02] Ironholds: :D I got started doing ops stuff after that, and now it’s a running gag because someone else in analytics did the same mistake me and milimetric did about 2 days after I started in ops... [13:35:23] same package even, I think [13:35:34] hah! [13:36:26] Ironholds: and equivalent machines! Mine was for stat1, and this time it was stat1003 I think [13:37:46] nice! [13:38:17] mforns, anyway. So, the easiest way to expose this and everything else is if we can get the architectural change merged - https://gerrit.wikimedia.org/r/#/c/197296/ [13:38:30] aha [13:38:34] I think nuria is pretty much happy with it, which is a good sign because Nuria sent me an email before she CRd my first patch [13:38:42] apologising in advance for being much stricter than otto ;p [13:38:51] and, well, she is :P [13:39:28] so once that's in I can throw in the second patch and you can CR if you're comfortable CRing and need it quickly. Shouldn't be hard, I just need to get the unit tests set up; nothing doesn't get unit tests [13:39:43] (except for stringContains and matchesPattern, because if those break, so do all the things with unit tests) [13:40:24] Ironholds, I see [13:40:35] ok, I'll bring this up in our stand-up [13:40:41] cool! [13:40:44] * Ironholds thumbs up [13:40:51] it's weird people ask me for stuff involving code these days [13:40:56] at some point it will get non-weird, I'm sure [13:41:07] :] [13:44:34] hokay, going afk for a bit. Back when everyone is awake :). mforns, LMK the standup outcome! [13:44:52] Ironholds, sure [13:55:15] Analytics, Analytics-EventLogging, Analytics-Kanban, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131817 (Nuria) Sorry I did not send e-mail to team yesterday. The root cause is a huge volume of events that are invalid that we can... [14:11:39] Analytics-Kanban, VisualEditor: VE events need to be sampled - https://phabricator.wikimedia.org/T93201#1131831 (Nuria) NEW a:Milimetric [14:15:17] Analytics-Cluster, Analytics-Kanban: Mobile PMs has reports on session-related metrics from Wikipedia Apps - https://phabricator.wikimedia.org/T86535#1131849 (mforns) a:mforns [14:17:02] Analytics-EventLogging, Analytics-Kanban, VisualEditor: VE events need to be sampled - https://phabricator.wikimedia.org/T93201#1131850 (Nuria) [14:18:05] Analytics, Analytics-EventLogging, Analytics-Kanban, operations: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1131852 (Nuria) Ticket: https://phabricator.wikimedia.org/T93201 [14:27:04] Analytics-EventLogging, Analytics-Kanban, VisualEditor: VE events need to be sampled - https://phabricator.wikimedia.org/T93201#1131880 (kevinator) p:Triage>High [14:28:23] Analytics-EventLogging, Analytics-Kanban, VisualEditor: VE events need to be sampled {lion} - https://phabricator.wikimedia.org/T93201#1131881 (kevinator) [14:29:47] Analytics-Kanban, Analytics-Wikimetrics: Utf-8 names on json reports appear as unicode code points: "\u0623\u0645\u064a\u0646" - https://phabricator.wikimedia.org/T93023#1131895 (kevinator) p:Triage>Normal [14:37:36] Analytics, Analytics-Kanban: Onboard Madhu - https://phabricator.wikimedia.org/T92985#1131920 (kevinator) p:Triage>Normal [14:40:03] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1131931 (kevinator) [15:27:55] nuria: want to deploy EL with me today, or should we wait? [15:28:02] ottomata: we cannot [15:28:14] ottomata: we need to deploy 1st the throttle for invalid events [15:28:49] you making changes to EL code to deploy? [15:28:51] sucheta, I forgot to give you your gift! [15:29:03] * Ironholds clearly now needs to go to LP [15:29:53] ottomata: otherwise we are tiping over .... [15:30:09] ottomata: but i will go from currently deployed changeset [15:30:13] ottomata: not master [15:30:26] ottomata: makes sense? [15:31:11] nuria, wait, what is tipping over? [15:31:14] just the error logs, right? [15:31:17] upstart logs are filling up? [15:31:18] ottomata: El itself [15:31:22] ottomata: yess [15:31:27] that is not EL, that is disk. [15:31:49] ottomata: well if no disk -> process shut down as they cannot log -> el no process events no more [15:32:02] ottomata: yesterday's log was 50G [15:32:10] ottomata: for validation failures [15:32:13] so delete more logs? [15:33:09] this one, right? [15:33:10] eventlogging_processor-client-side-events.log [15:33:14] growing at about 2MB / sec [15:33:25] ottomata: yes, i need to not log every error [15:34:10] ottomata: as we do not need a 50G log every day as it's going to fill machine faster than we can clean it up [15:34:24] nuria, why not just not log the whole event in the error logs? [15:34:26] just the error [15:35:21] or, since this is kind of an emergency stop gap, just logrotate every hour or something [15:35:50] ottomata: I think we should log less, but with enough info so errors are actionable [15:36:03] ok so [15:36:59] ottomata: so per mforns idea we could random sample errors [15:37:10] nuria [15:37:10] here [15:37:11] https://github.com/wikimedia/mediawiki-extensions-EventLogging/blob/master/server/bin/eventlogging-processor#L70 [15:37:14] just remove raw_event [15:37:17] from the log line. [15:37:28] or no [15:37:29] i'm sorry [15:37:29] hm. [15:37:31] no [15:37:32] ottomata: but then errors are not actionable [15:37:35] the stuff in (%s) is the problem [15:37:54] ottomata: the errors we log should have enough info to fix the issue [15:38:02] i'm looking at that file now [15:38:13] each message has the full schema in it [15:38:17] with descriptions of each field [15:38:26] ottomata: ya that is redundant [15:38:31] that is why the file is so huge [15:38:35] yes there are lots of errors [15:38:38] but each line is showing that [15:39:10] ottomata: right, i will remove that. [15:41:39] hm, i think it must be coming from exception thrown by jsonschema.Draft3Validator .validate [15:43:22] you should maybe just catch different types of errors and do things differently [15:44:20] hmmmmmMMMMm [15:44:34] nuria, i don't know python exceptions well, if you print e.message, rather than e. it might do it for you. [15:44:58] ottomata: i will get a codechange as soon as we are out of meeting [15:45:06] ottomata: will add you CR [15:47:11] ottomata: let me use beta labs for a bit cause i can test it there [15:48:31] ja, nuria, i just tried it [15:48:34] i tworks [15:48:35] it works [15:48:49] i can submit patch now, shall I? [15:50:41] it is happening! :) [15:52:06] nuria: https://gerrit.wikimedia.org/r/#/c/197929/1/server/bin/eventlogging-processor [15:52:30] i ididn't commit it on top of my changes, but it i did do it on master...i probably will ahve to put it on its own branch if you want to deploy that without my changes [15:52:34] rigiht? [15:56:56] ottomata: I think that is not going to change matters much, exceptions in python are not like the ones in java [15:57:26] nuria [15:57:29] i tried it, it does. [15:57:30] ottomata: but yes, patch needs to be on top of latest deployed changeset, let me test that in beta-labs [15:57:35] https://github.com/Julian/jsonschema/blob/master/jsonschema/exceptions.py#L62 [15:57:44] if you print the message [15:57:55] it will print everything from that __unicode__ method [15:58:00] which includes the schema [15:58:13] , but you can just print e.message, which is just the error message [15:58:36] ottomata: ah, ok, if you tried then sounds good, let me transplant that [15:58:36] then, in the parens I get [15:58:37] (u'editingSessionId' is a required property) [15:58:44] rather that that HUGE string about everyhting [15:58:55] ottomata: everything and its mom [15:58:59] haha [15:59:03] ottomata: ok, let me transplant [15:59:22] nuria, may it be that my patch that you merged yesterday is causing these issues? [15:59:30] mforns: NOOOOO [15:59:44] mforns: your patch, if anything, reduces throughput from js client [15:59:53] Analytics, Analytics-Kanban: Turn off WP Zero's Limn-Dashboards & put up a "moved sign" - https://phabricator.wikimedia.org/T92920#1132184 (kevinator) Desired outcome: 1- User goes to http://gp.wmflabs.org/dashboards/ncell-nepal 2- User sees message "This Wikipedia Zero dashboard is no longer available...... [15:59:54] mforns: do not worry [16:00:24] nuria, oh again I forgot that the server-side error log, was only for the dev-server [16:00:33] mforns: right [16:04:45] ottomata: this is same on top of current latest deploy: https://gerrit.wikimedia.org/r/#/c/197930/ [16:05:54] nuria, mine was on otp of the latest deploy too [16:05:56] it is the same [16:06:03] ottomata: ahhh sorrry [16:06:05] https://gerrit.wikimedia.org/r/#/c/197929/ [16:06:17] but, i'm not sure how to deploy, cause gerrit wants to rebase it [16:06:20] i think we need to make a branch [16:06:25] or, just deploy it as a patch manually [16:06:33] ottomata: k, you want to deploy it? [16:06:38] would love to ! :) [16:06:43] ottomata: nbo , we deploy on box [16:06:58] ok, so pull from gerrit and deploy? [16:07:00] on /srv/deployment/evenetlogging/Eventlogging [16:07:01] or just edit on box? [16:07:03] and deploy? [16:07:07] ottomata: i do not think there is another way [16:07:13] ottomata: pull from gerrit [16:07:15] we could make a real branch [16:07:26] ottomata: as git-deploy doesn't let you eploy non merged stuff [16:07:27] but, if you want to deploy the kafka thing with me today...then we don't have to do that [16:07:27] :) [16:07:34] right, if we made a branched and merged it there [16:07:44] ottomata: call me grandma but i rather fix one thing at a time [16:07:50] oh indeed [16:07:51] i'm just saying [16:07:58] ottomata: jaja [16:07:59] if we plan on doing a real deploy later, then we can manuallay make this change now [16:08:13] ottomata: let's fix the logging issues 1st [16:08:15] but, if you want to wait a while on that, then we might want to do a real deploy rather than a manual change for this [16:08:36] ottomata: I rather fix the acute problem with manual deploy, we will log it on sal [16:08:40] ok [16:08:48] ottomata: i can do it [16:08:48] just on vanadium, right? [16:08:53] ottomata: yes [16:09:02] naw, show me how. [16:09:07] ottomata: k, [16:09:10] so, checkout this gerrit change, and then python setup install? [16:09:24] ottomata: go to /srv/deployment/eventlogging/Eventlogging and pull from gerrit [16:09:42] ottomata: git fetch https://gerrit.wikimedia.org/r/mediawiki/extensions/EventLogging refs/changes/70/188270/6 && git checkout FETCH_HEAD [16:09:47] ottomata: you will need sudo [16:09:52] k [16:10:14] ottomata: above is an example [16:10:18] ottomata: not real thing [16:10:18] yes [16:10:25] ottomata: then we do [16:10:28] then python setup.py install? [16:10:42] ottomata: git checkout -b changeset [16:10:50] i did that in the same command [16:10:51] like [16:10:55] ottomata: so we know what piece of code we have there [16:11:00] git fetch https://gerrit.wikimedia.org/r/mediawiki/extensions/EventLogging refs/changes/29/197929/2 && git checkout -b error-log-reduction FETCH_HEAD [16:11:00] ottomata: k you are a pro [16:11:11] ottomata: then we do the python install thingy [16:11:17] in server directory [16:11:19] ja [16:11:21] ready for that? [16:11:24] ya [16:11:43] done [16:11:44] ottomata: and then [16:11:56] eventloggingctl stop [16:12:02] eventloggingctl start [16:12:02] the whole thing?! [16:12:06] ok [16:12:20] doing it. [16:12:39] ottomata: and later ps -auxfw [16:12:45] to make sure all started nicely [16:12:50] cool, error logs much smaller now [16:12:53] lets see how much... [16:12:54] ottomata: and tail -f on /var/log/upstart [16:13:29] ok, looks like about 300KB/s now [16:13:33] vs 2MB / sec [16:13:35] before [16:13:45] ottomata: BIG improvement ... pufffff [16:13:59] nuria, btw, do you know pv? [16:14:01] ottomata: lemme look [16:14:01] pipe viewer? [16:14:02] really cool [16:14:10] tail -f /var/log/upstart/eventlogging_processor-client-side-events.log | pv > /dev/null [16:14:22] ottomata: looks good [16:14:33] ottomata: so let's merge this same change to master [16:15:21] ottomata: anbd let's log to sal what we deployed and that's it (for now) BIG thanks! [16:17:25] !log manually deployed change to eventlogging on vanadium that reduces size of error log messages in /var/log/upstart: https://gerrit.wikimedia.org/r/#/c/197929/ [16:18:50] ottomata: super thanks again [16:19:51] ja you know, just being selfish! I wanna deploy! [16:20:08] nuria: we shoudl just rebase and merge this onto master then [16:20:09] https://gerrit.wikimedia.org/r/#/c/197929/ [16:20:16] yes [16:20:30] so the "kafka" changes carry these too. [16:20:48] rebase was fine, i'll let you merge it :) [16:26:35] Analytics, Analytics-EventLogging, Analytics-Kanban, operations, Patch-For-Review: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1132304 (Nuria) Crisis averted, loogs are growing to 300kbs sec not to *ahem* 2MB per sec. Resolving ticket. [16:26:49] Analytics, Analytics-EventLogging, Analytics-Kanban, operations, Patch-For-Review: Disk space full on vanadium from logs in /var/log/upstart - https://phabricator.wikimedia.org/T93185#1132307 (Nuria) Open>Resolved [16:30:16] nuria: so yeah ujmmmm deploy today? circle one: yes/no [16:30:16] ? [16:30:20] :D [16:30:46] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132318 (csteipp) Hi all, I looked through this again and wanted to get some clarification on it. When I talked wit... [16:30:58] ottomata: yes, noon PST? I have two meetings until then [16:31:06] ottomata: is that too late? [16:31:59] hm, nuria, i feel comfortable doing it on my own, but if there was a problem i'd want you around. noon is 3 pm my time, i think that should be fine, right? [16:32:08] even if i have to go, you will be around for a bit more? [16:32:22] ottomata: quite a bit more , like 6/7pm PST [16:32:28] so 6 more hours [16:32:32] ok, then that's fine, let's do it then. [16:32:40] k [16:36:36] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1132335 (mforns) NEW [16:49:56] Analytics-Engineering, Analytics-Wikimetrics: Unable to add a custom cohort user - https://phabricator.wikimedia.org/T91751#1132374 (kevinator) The documentation was anonymously edited today with a better description of the steps to follow: https://www.mediawiki.org/wiki/Analytics/Wikimetrics/Help#Adding_... [17:17:41] Ironholds, I get a gift? YAY. [17:18:06] sucheta, I found a second {{citation needed}} stamp [17:18:57] Oh, you're the best! I want it. [17:23:49] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1132492 (Milimetric) The Rolling New Active Editor metric did not complete for this cohort with default parameters. ``` [2015-03-19 16:18:46,078: INFO/Worker-29] wikimetrics.models.rep... [17:52:09] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1132599 (dr0ptp4kt) [17:53:33] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1063059 (dr0ptp4kt) [17:56:50] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1132651 (dr0ptp4kt) a:dr0ptp4kt [18:00:15] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132687 (Nuria) >When I talked with Aaron about setting the cookie, it was going to be a generic cookie for the mont... [18:04:55] Analytics-Cluster, Analytics-Kanban, Performance: Implement Last-Access cookie [34 pts] {bear} - https://phabricator.wikimedia.org/T88813#1132716 (Nuria) @RobLa-WMF: >Ok, it took me a little bit to wrap my head about what y'all are doing with this, and was kind of skeptical. Assuming I understand this... [18:05:12] Analytics-Cluster, Analytics-Kanban, Performance: Implement Last-Access cookie [34 pts] {bear} - https://phabricator.wikimedia.org/T88813#1132717 (Nuria) [18:17:11] Analytics-Kanban, Patch-For-Review: Failure Types by User Type - https://phabricator.wikimedia.org/T91123#1132784 (Milimetric) Note: this and all other analyses done this quarter for the Edit schema are no longer valid. This is because we have an unknown number of invalid events in our logs that affect th... [18:31:13] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132835 (kevinator) [18:31:14] Analytics-Cluster, Analytics-Kanban: Add processed user agent to refined tables - https://phabricator.wikimedia.org/T91793#1132837 (kevinator) [18:31:15] Analytics-Cluster, Analytics-Kanban: Refine webrequest x_analytics field into a map in the refined table. - https://phabricator.wikimedia.org/T89396#1132836 (kevinator) [18:31:43] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132841 (Anomie) >>! In T92977#1132687, @Nuria wrote: > Cookie lives forever, but its value is changing on every acc... [18:31:52] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132852 (kevinator) [18:31:54] Analytics-Cluster, Analytics-Kanban, Epic: {epic} WMF has UC report per project per month & day {bear} - https://phabricator.wikimedia.org/T88647#1132851 (kevinator) [18:34:28] Analytics-Cluster, Analytics-Kanban: Add user_agent map to refined tables - https://phabricator.wikimedia.org/T91793#1132872 (kevinator) [18:35:01] is there a reason this and related tasks are on the Wikimetrics board? https://phabricator.wikimedia.org/T76761 [18:36:28] fhocutt, kevinator_ put it there, although maybe it was by accident? [18:36:47] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132875 (Nuria) @Anomie Let us know if there are privacy concers but per comments above we do not think there is any... [18:37:09] it doesn't seem to be related, but I don't want to remove it if there was a reason it was there [18:37:42] Analytics-Cluster, Analytics-Kanban: Add x_analytics map to refined tables - https://phabricator.wikimedia.org/T89396#1132880 (kevinator) p:Low>Normal [18:38:36] fhocutt, maybe it has something to do with wikimetrics because it depends on productionizing the scheduler part of it, but I'm not sure [18:39:03] fhocutt: I'm looking at the task again [18:39:40] thanks kevinator_ [18:40:44] fhocutt: yeah, this task should not be in the wikimetrics project [18:40:55] I will move it now and any of its children [18:41:08] thanks! [18:42:37] Analytics-Cluster, Analytics-Engineering: Mondrian has access to the PostgreSQL database - https://phabricator.wikimedia.org/T76766#1132895 (kevinator) [18:42:51] Analytics-Cluster, Analytics-Engineering: make cubes as windows into warehouse - https://phabricator.wikimedia.org/T76767#1132896 (kevinator) [18:43:20] Analytics-Cluster, Analytics-Engineering: EPIC: Aggregating and Shaping the data for browsing - https://phabricator.wikimedia.org/T76761#1132904 (kevinator) [18:44:35] Analytics-Cluster, Analytics-Engineering: Mondrian has access to the PostgreSQL database - https://phabricator.wikimedia.org/T76766#819533 (kevinator) test [18:45:33] Analytics-Cluster, Analytics-Kanban, Performance: Implement Unique Clients report on cluster using x-analytics header & last access date {bear} - https://phabricator.wikimedia.org/T92977#1132925 (Anomie) Both tasks may refer to the same document, but the //discussion// on the other ticket clearly doesn... [18:50:27] Analytics-EventLogging, Analytics-Kanban, VisualEditor, WMF-NDA: Wikitext Eventlogging events not validating - https://phabricator.wikimedia.org/T93242#1132939 (Nuria) NEW [18:51:28] ottomata, yt? [18:52:24] Analytics-EventLogging, Analytics-Kanban, WikiEditor, WMF-NDA: Wikitext Eventlogging events not validating - https://phabricator.wikimedia.org/T93242#1132948 (Catrope) [18:55:14] mforns: ja [18:55:23] hey! [18:55:35] ottomata, are you planning to work on the x-analytics udf? [18:55:59] because if you are on other things, maybe I can do that? [18:57:20] ? [18:57:25] oh the xanalytics map thing? [18:57:30] joal was going to work on it [18:57:37] but ya know, baby happened [18:58:05] https://phabricator.wikimedia.org/T89396 [18:58:16] i think he may have already done some work on it [19:03:03] but mforns i dunno if he has or not [19:03:12] hmm, you looking for things to do? [19:03:13] : [19:03:14] :) [19:03:36] ottomata, no it's because I need it for the other task [19:03:38] mforns: i think joseph has done almost done [19:03:57] mforns: email joseph and ask if he can submit what he has, and we can finish it up [19:04:07] mforns: but maybe if you send him an e-mail you can verify [19:04:15] ah yes sorry [19:04:22] nuria, ottomata: ok, sure [19:04:22] mforns: for now you can just use the udf to extract fields, no? [19:04:29] sorry, or not udf [19:04:30] function :) [19:04:31] umm [19:05:02] https://github.com/wikimedia/analytics-refinery-source/blob/master/refinery-core/src/main/java/org/wikimedia/analytics/refinery/core/Webrequest.java#L69 [19:05:12] ottomata, yes, sure I'll use temporary code until the udf is merged, and send an email to joal [19:05:28] ooohh [19:05:33] ok, thanks! [19:05:55] yup :) [19:06:03] nuria: ya wanna? [19:06:15] ottomata: yes, let's do it [19:06:41] batcave? [19:06:41] yeehaw [19:06:42] k [19:11:39] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1133060 (Milimetric) Even worse, the report appears successful and has a result!! https://metrics.wmflabs.org/reports/result/6da4d2f5-16d7-4285-b5eb-856126fc9833.json [19:11:50] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1133061 (Milimetric) [19:12:41] wow ^ [19:21:32] milimetric, I saw your email, have you found a particular issue in the infrastructure? [19:22:02] mforns: yeah, the fact that we're running huge queries on top of views on a shaky database [19:22:03] :) [19:22:07] but nothing new, no [19:22:15] milimetric, ok [19:26:48] x-analytics map should be pretty trivial [19:27:05] "split on commas, for each split_string, split on equals, first value is key, second value is value, reassemble" [19:48:11] Analytics-Engineering: Pageviews definition undercounts app requests - https://phabricator.wikimedia.org/T93255#1133116 (Ironholds) NEW [20:20:29] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1133211 (kevinator) p:Triage>High [20:34:38] nuria, ping :) [20:46:03] Ironholds: hello [20:46:08] heyo :) [20:46:09] Ironholds: i own you a CR [20:46:28] well, nobody owes me nuthin'! But if you could do it I can get some work done for mforns and updating the PV def :) [20:46:37] Ironholds: qq: do you know how many PVs and uniques we got over the course of the past year? (either full calendar year or past 12 months -- whichever) [20:47:02] ori, uniques is unpossible, but pageviews...let me get my calculator. Including or excluding bots? [20:47:15] excluding [20:47:25] s/calendar/the SQL table I dumped things into [20:47:27] bots view pages? [20:47:31] Analytics-Kanban, Analytics-Wikimetrics: Troubleshoot Wikimetrics RAE reports - https://phabricator.wikimedia.org/T93217#1133276 (kevinator) @milimetric, can you clarify: The title refers to RAE, but the comment says RNAE didn't complete. Are all rolling metrics not working? [20:47:49] bots view pages. Bots tend to be run by assholes; the legacy definition didn't even TAG them [20:48:04]