[00:14:01] Eyyyyyy Snaps_! [00:15:04] what does log.syslog = true log to? [00:15:23] like, if I want to put the varnishkafka logs in their own file [00:15:35] do I need to add a (r)syslog thing [00:15:38] ? [00:15:55] what facility? [02:31:05] (PS1) Milimetric: timeseries checkpoint [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84706 [02:32:18] (CR) Milimetric: [C: 2 V: 2] "only working for namespace edits right now, but good progress" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84706 (owner: Milimetric) [05:48:56] (PS2) Milimetric: timeseries working in bytes_added and edits [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84706 [05:49:18] (CR) Milimetric: [C: 2 V: 2] timeseries working in bytes_added and edits [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84706 (owner: Milimetric) [13:34:22] (PS1) QChris: Remove "global_south_editor_fractions" [analytics/global-dev/dashboard-data] - https://gerrit.wikimedia.org/r/84746 [13:34:52] yo qchris [13:34:57] Hi drdee. [13:35:10] can you demo #1079 today? [13:35:10] drdee hopes that someone will have a look at https://mingle.corp.wikimedia.org/projects/analytics/cards/1079 [13:35:34] don't you love CommenderData? [13:36:00] He is great? Why shouldn't I love him :-P [13:36:05] Our hardware is too slow to cope with demand [13:36:08] Hehe. [13:36:20] s/?/!/ [13:36:28] that's a reason :) [13:36:35] I am not sure how to demo monitoring :-/ [13:36:56] right, well maybe show the chart [13:37:01] and explain how monitoring works [13:37:01] I can show that all graphs get updated. [13:37:11] and yes show that as well [13:37:19] Ok. [13:37:24] shall i move 1079 to ready for showcase? [13:37:40] average, around? [13:37:53] The monitoring needs some more tuning. I'll do that today. [13:38:08] k [13:38:21] but you can demo it right? [13:38:25] Yes. [13:40:44] (CR) Diederik: [C: 2 V: 2] "Go ahead." [analytics/global-dev/dashboard-data] - https://gerrit.wikimedia.org/r/84746 (owner: QChris) [13:41:11] Thanks drdee! [13:43:48] hey qchris, did I just miss a call from you? [13:44:16] milimetric: Mhmm? I do not think so. [13:44:38] milimetric: Hi :-) [13:44:41] google just sent me a message via hangouts [13:44:44] weird :) [13:45:01] By google machine is sitting on the shelf (turned off) [13:45:04] did you perhaps meditate about calling me? Google is able to detect that you know [13:45:09] But you never know with google. [13:45:20] Yes :-D [13:45:30] ok, cool, at least there's an explanation [13:45:33] let me know if you ned me :)( [13:45:39] morning milimetric [13:45:42] morning drdee [13:53:44] milimetric: any mingle cards that can be moved to showcase? [13:54:01] would you like to look over timestamps with me? [13:54:07] timeseries is ready for bytes added and edits [13:54:07] i sure would [13:54:24] did we figure out the new bat cave link? [13:54:27] nope [13:54:33] carp [13:54:55] https://plus.google.com/hangouts/_/e0ed8fd7dd064a696192931c8c882c6b8d8ecb21 [14:23:02] ottomata, in da house? [14:29:47] ok drdee, it works as long as I configure it properly [14:30:01] so I can demo as-is [14:30:12] but there just won't be any interesting data [14:30:19] unless we deploy to prod [14:30:44] which we could probably do if we turn off survivor metric or if stefan has finished it [14:30:54] i'm gonna go get breakfast [14:31:47] I have not finished it [14:32:16] you can turn it off in production, I'm still working on it. it's not going to be ready for the showcase in a few hours [14:34:41] cool, I'm not super sure timeseries will be ready either [14:34:43] :) [14:34:58] any trouble with date arithmetic or anything? [14:35:03] have you pulled my latest? [14:35:24] date arithmetic is ok [14:36:15] I am working on adding the artificial sunset DarTar explained in the e-mail [14:36:18] as well as tests to cover it [15:30:43] oh I missed your message average, you're right that was scope creep that I think we didn't deal with very well [15:31:09] we should communicate that to drdee and make sure we re-estimate the card accordingly [15:31:15] got it [15:31:38] we need to discuss this next time, we all should be more vigilant about this [15:59:30] (PS1) Stefan.petrea: New survivor metric [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84756 [16:00:32] (PS2) Stefan.petrea: New changes on survior metric [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84756 [16:04:12] ottomata: log.syslog and log.stdout is for varnishkafka's own debug/error logs. [16:04:25] ottomata: the facility is daemon [16:33:25] (PS1) Milimetric: pages created timeseries implemented [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84764 [16:33:37] (CR) Milimetric: [C: 2 V: 2] pages created timeseries implemented [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84764 (owner: Milimetric) [16:47:18] daemon, ok, thanks Snaps_. [16:52:09] (PS1) Milimetric: fixes sorting order [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84767 [16:52:22] (CR) Milimetric: [C: 2 V: 2] fixes sorting order [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84767 (owner: Milimetric) [17:03:16] (PS3) Stefan.petrea: New changes on survior metric [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84756 [17:07:15] ottomata: should probably rename them to something else not to confuse it with the other pesky detail of varnishkafka: the varnish log itself [17:09:50] oh hmmm [17:09:51] yeah [17:09:53] um [17:10:23] hmm, i dunno, i mean, i could tell what they were, just not how they worked (daemon facility) [17:10:51] not sure what a better name would be though, but hmm [17:18:42] okay! [17:18:43] good [17:18:44] then no need [17:18:57] Will just document the facility, maybe even make it [17:22:40] Snaps_: so if I wanted to log to /var/log/varnishkafka.log [17:22:42] I'd add a [17:22:59] daemon.* /var/log/varnishkafka.log entry in rsyslog.d/ somewhere? [17:27:04] I dont know rsyslogd details, but that sounds about right. But you'll also get all other daemon stuff in there. [17:28:11] yeah [17:28:22] i'd like to isolate them, not sure how this is usually done [17:30:58] http://serverfault.com/questions/288132/using-rsyslog-to-create-different-log-files-for-different-processes [17:35:10] hey average, I'm back from lunch [17:35:25] i'm not sure why all the tests pass for me but not you [17:35:26] Did you check for a virus on your system? [17:35:31] lol [17:35:44] ROFLOL [17:35:53] that's the best response! [17:36:00] :-)) [17:36:29] hahhahahahaaaaa [17:37:09] Snaps_: haha, i'm looking at that right now :) [17:37:48] hey drdee, timeseries is ready to showcase, deployed [17:37:59] but i put both that and timestamp in sign-off [17:38:05] A-W-E-S-O-M-E [17:38:07] since we don't technically have sign-off from anyone externam [17:38:10] *external [17:38:46] oh, never mind, the wall is all crazy like :) [17:38:54] it's the next sprint [17:39:01] yea [17:39:15] i'm not sure what happened when i moved the card before refreshing [17:39:29] but we could mark them done after today hopefully and it won't matter [17:39:37] k [17:39:47] the only problem I can see is that arwiki is unreachable from labs [17:39:53] I asked in labs but probably nobody's in yet [17:47:37] tobyyyyy [17:47:38] where are you? [17:56:07] sprint demo hangout: https://docs.google.com/a/wikimedia.org/presentation/d/1k-ookW_C8wUeA3ut0aRJXeYwmh8_EEuJv3RH0D8n-8E/edit#slide=id.p [17:56:12] i mean https://plus.google.com/hangouts/_/8adb135d7d3e3ee4fc21f5115f9240abc328a48e [18:00:21] milimetric, ottomata, qchris, average: ^^ [18:01:40] working on it [18:01:43] i'm here in person this time, woooooo [18:02:46] cool [18:02:53] what's happening with the hangout [18:02:55] ? [18:04:40] iunnnooo [18:04:55] working on it [18:12:50] link please [18:12:58] ottomata the link please [18:13:20] https://plus.google.com/hangouts/_/bb0f03183c1074ac4432664543a6abd24c39d2fe?hl=en [18:13:24] that's the link [18:13:46] please mute [18:14:59] I do not like to get recorded or streamed live to the internet :-( [18:15:24] drdee, milimetric: Could you briefly explain the geowiki stuff? [18:15:29] (Instead of me) [18:15:39] you sure qchris? [18:15:43] * average is not allowed [18:15:45] I could go over it, definitely [18:15:53] qchris: then don't check the checkbox [18:15:54] drdee / ottomata, invite average [18:15:55] milimetric: That would be awesome. Thanks! [18:15:58] :P [18:16:10] ok, no prob. qchris [18:16:22] qchris: can we turn off the cam ? [18:17:07] average: You cannot even join the hangout when you do not check box at the "I agree ..." text [18:17:24] average: Have fun withouth me then. [18:17:37] average: invited you [18:17:53] qchris, is there any place I can look at the monitoring output of the script? [18:18:02] qchris: cmon join [18:19:02] milimetric: It's a cronjob, so it's sent via email to the stats user. [18:19:35] ottomata: can you invite qchris as well pls ? [18:20:25] milimetric: So I would have just shown the graphs on http://gp.wmflabs.org [18:20:53] The tabs core & Brazil are covered by geowiki [18:21:02] (grants is not) [18:22:23] qchris: you can join the hangout [18:22:24] and all the per-wiki graphs. So eg. for dewiki: http://gp.wmflabs.org/graphs/de_top10 [18:22:27] https://plus.google.com/hangouts/_/stream/bb0f03183c1074ac4432664543a6abd24c39d2fe?pqs=1&hl=en&authuser=0 [18:23:05] qchris: ^^ [18:23:12] it's hard to read the code [18:24:18] Finally I would have shown https://wikitech.wikimedia.org/wiki/File:Geowiki_workflow_2013-09-15.png [18:24:30] and https://wikitech.wikimedia.org/wiki/Analytics/Geowiki [18:24:33] qchris: i invetd you [18:24:35] at that other email addy [18:24:47] Thanks ottomata. [18:25:22] ottomata: I will not join you. mil imetric will take over my part. [18:26:00] ok oo [18:26:02] coo [18:34:00] qchris: you can come in now, we stopped recording [18:34:45] https://plus.google.com/hangouts/_/bb0f03183c1074ac4432664543a6abd24c39d2fe [18:34:47] ja done recording [18:34:50] no more broadcasting [18:34:52] show yourself! [18:35:22] Ok :-D [18:35:35] qchris you can join us [18:36:06] drdee: That hangout again says it's recorded :-( [18:36:16] but it's not [18:36:28] hmmm [18:36:31] well it used to be [18:36:35] but it isn't broadcasting anymore [18:36:37] "I agree that my participation in this Hangout On Air will be reporded ...[...]" [18:36:41] mabye we should just switch? [18:37:15] ok do this one: [18:37:15] https://plus.google.com/hangouts/_/472235473bf08b42f0039f2d77f3d8e36a0d3952 [18:37:45] ottomata: That's better :-D [18:37:54] ottomata: Could you add me other address? [18:38:20] No checkboxes I need to click on :-) [18:40:14] https://plus.google.com/hangouts/_/472235473bf08b42f0039f2d77f3d8e36a0d3952 [18:40:19] milimetric: ^ [18:41:15] milimetric: ^^ [18:41:22] i joined! [18:41:29] wtf it kicked me out [19:12:39] If you guys are hunging out still I'd like to join too [19:14:47] OrenBochman: you mean the analytics showcase? [19:14:59] we have it every 2 weeks at 11AM PST [19:15:03] but [19:15:05] we taped this one! [19:15:09] ottomata: what's the link? [19:15:26] uhhhh [19:15:54] http://www.youtube.com/watch?v=vRAMBk0EBkg [19:16:01] timestamps! [19:16:13] i need help! [19:16:47] hi SarahStierch [19:16:51] Hey OrenBochman [19:17:42] I'm a little confused about the timestamp stuff. I can't quite crack the nut on making it work. [19:18:12] I've created a small app for visualising neutrality of artilces [19:18:19] Interesting! [19:18:36] Is it online? [19:18:38] I'm now adding a second tool to check sexist/femenist language [19:18:39] Stupid question [19:18:43] interesting [19:18:48] it is not online [19:18:54] It needs a special server [19:18:57] especially with the whole hub hub around the "storming" of wikipedia by feminists or whatever [19:18:58] ah ok [19:19:04] keep me posted on when it is [19:19:11] so can anyone help me about this timestamp thing on wikimetrics? [19:19:32] what about changing End - 2013-03-09 00:00:00 [19:19:36] to End - 2013-03-09 23:59:59 [19:19:41] hi SarahStierch [19:19:45] hey milimetric [19:19:46] gah [19:19:48] that's a pain in the ass [19:19:49] ok [19:19:50] * SarahStierch tries [19:19:52] I'm the wikimetrics dude :) [19:20:12] oh yeah, pain in the butt [19:20:25] i want to pull data on a four hour edit-a-thon [19:20:26] any femenist pov needs creating a femenist lexicon [19:20:50] drdee: there is one thing I forgot to mention: Our Bat cave died during this spring! Could we get Batman a new Batcave to put all his shiny gear in? [19:21:09] yes we need one but we need to figure out a way to get permanent links working again [19:21:17] Should I chose a time series? [19:21:22] so SarahStierch, per DarTar's instructions, the left part of the interval is Not included in the results [19:21:24] I'm a bit confused by that [19:21:27] this is definitely annoying [19:21:50] yeah, I demo-ed it at the sprint demo just now, but let's go over it in a hangout [19:21:55] I have an interview for the next hour or so [19:21:57] but after that? [19:22:03] dammit i'm still getting that error [19:22:07] [object Object] on every page [19:22:36] SarahStierch: that basically means the server had an error and it's saying Object object because it doesn't want to barf on your pretty clothes :) [19:22:43] it does it all the time [19:22:47] for two days i've had Object object [19:22:52] ALL THE ERRORS! [19:22:53] yea, it just happened because we enabled https [19:22:53] and yes no barfing [19:22:57] * SarahStierch thanks wikimetrics for not puking on her [19:23:00] ah ok [19:23:16] so the https is wreaking havoc and we haven't had time to look at it because we're working on timeseries [19:23:22] not to worry, all will be ok soon enough :) [19:23:22] damnit [19:23:28] so [19:23:30] ignore that damnit [19:23:36] like I said, after the interview, let's talk? [19:23:39] YES OK IT WORKED! [19:23:44] how come wikimetrics won't let me log in with my SUL login [19:23:54] interview? [19:24:09] I mentioned above I have to interview a nice gentleman for the analytics team [19:24:19] I'll be done at 16:30 our time [19:24:20] ah ok! [19:24:24] so 13:30 yours [19:24:30] wait what do we need to talk about? :) [19:24:37] you said you were having a problem [19:24:43] with timestamps [19:24:46] drdee provided me the answer [19:24:47] and dammit dammit, etc [19:24:47] :) [19:24:51] BOOEM [19:24:53] oh, haha [19:24:56] :) [19:25:00] hm it crashed [19:26:20] not the best bug report I've ever gotten OrenBochman ^^ [19:26:22] :P [19:26:32] SarahStierch: happy? [19:26:34] YES [19:26:43] great! [19:26:45] this is making my epic evaluation on the productivity of edit-a-thons even ore... [19:26:49] more….depressing [19:26:50] not the best error message I ever got [19:27:11] …depressing, why? [19:27:12] Management insisted we wouldn't need to waste our time writing unit tests [19:27:23] aarghghhhh [19:27:24] so far my research has been showing that they aren't that productive [19:27:30] right [19:27:32] mmmmmm bummer [19:27:48] well they are not designed for productivity [19:28:04] i'm doing a self evaluation of my own events, and then doing a broader evaluation of English Wikipedia edit-a-thons. Just to see where the investment is…and why people even do them. [19:28:04] It can't be broken, it passes all unit tests [19:28:18] Sure, they're social and fun, but..i can't make assumptions [19:28:21] perhaps some are more productive than others [19:28:26] and many don't seem successful at retaining new editors [19:28:58] a lot of this is stuff i've always guessed about, but seeing the numbers is really interesting [19:29:03] wikimetrics is my new favorite thing on earth [19:29:59] my point is people are doing hackathons for a number of reasons. Retention has never been a top level concern [19:30:32] so going back and testing for this is a strange methedology [19:30:51] many wikipedians like doing edit-a-thons to retain and train new editors [19:30:53] i have done it [19:30:55] many many times [19:31:08] train is not retain [19:31:12] ugh [19:31:27] its like a company making a sale and making multiple sales [19:31:33] ughhhh [19:31:36] it all just depends [19:31:43] and me using the word "train" was a fail i guess [19:31:44] nevermind [19:31:55] i have things to do and i'm not here to get meta about this right now :) [19:32:06] sure [19:32:42] we had debates on this in the chapter meeting a couple of days ago [19:33:09] "OrenBochman: how come wikimetrics won't let me log in with my SUL login" [19:33:25] we do not support SUL yet, you will need to use a google account for now [19:33:39] yep how come you are asking users to identify via email ? [19:34:20] is that legal? [19:34:29] of course it's legal :) [19:34:30] we will support Mediawiki OAuth soon but we did not want to write an entire login / password reset functionality and then through it away [19:34:37] through => throw [19:35:02] is this because it's running in labs ? [19:35:07] is what? [19:35:27] OrenBochman: are you refering to wikimetrics ? [19:35:34] yes [19:35:34] we are waiting for stable Mediawiki Oauth support [19:35:40] me too [19:36:13] in fact I've abandoned my main project [19:36:15] great! [19:36:55] we missed the opportunity for this year [20:32:41] heya Snaps_, how hard would it be to add a log.file config property? to make logs go to a file instead of stderr/syslog [20:32:41] ? [20:32:50] it would need SIGHUP support for logrogate [20:33:43] not very hard at all [20:34:09] no luck with rsyslogd? [21:02:10] (PS1) Milimetric: fixing the default filling and slice problem [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84862 [21:02:20] (CR) Milimetric: [C: 2 V: 2] fixing the default filling and slice problem [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/84862 (owner: Milimetric) [21:03:07] its working, but the rotate is not ideal at all, and its kinda weird with file permissions [21:23:32] ottomata: okay, its not a problem to add, but you guys need to decide on it [21:24:23] yeah, paravoid seemed like he was leaning towards adding it, lemme double check with him [22:02:29] (PS1) QChris: Add 'Active Editor Totals' [analytics/geowiki] - https://gerrit.wikimedia.org/r/84875