[00:36:32] (PS1) Nuria: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 [01:04:11] (CR) Milimetric: Add wikimetrics api and data converter (1 comment) [analytics/dashiki] - https://gerrit.wikimedia.org/r/156453 (owner: Milimetric) [01:04:28] (CR) Milimetric: [C: -1] "I disagree with jquery in the global scope." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156453 (owner: Milimetric) [09:28:58] Analytics / General/Unknown: Constantly increasing number of fundraising processes on gadolinium - https://bugzilla.wikimedia.org/70053 (christian) [09:29:00] Analytics / General/Unknown: Webstatscollector's pagecounts file shaky between 2014-08-24 14:00 and 2014-08-27 21:00 - https://bugzilla.wikimedia.org/70118 (christian) NEW p:Unprio s:normal a:None TL;DR: For the files between 2014-08-24 14:00 and 2014-08-27 21:00 webstatscollector output ar... [09:30:14] Analytics / General/Unknown: Constantly increasing number of fundraising processes on gadolinium - https://bugzilla.wikimedia.org/70053#c4 (christian) NEW>RESO/FIX Closing the bug, as the processes are now gone. Thanks ottomata! [11:06:14] Analytics / General/Unknown: Webstatscollector's pagecounts file shaky between 2014-08-24 14:00 and 2014-08-27 21:00 - https://bugzilla.wikimedia.org/70118#c1 (christian) (In reply to christian from comment #0) > [...] restarted on 2014-08-28 ~15:32. That timestamp is meant to read "2014-08-27 ~15:32.... [14:59:01] springle: SAL says you restarted morebots yesterday. Since morebots also gave us an "analytics-logbot" for this channel (which is gone since), do we need to do anything to get it back, [14:59:16] or does it just take long to come back up and join us in the channel? [15:00:01] s/yesterday/before I got up/ :-) [15:01:14] qchris: i only touched the #wikimedia-operations bot. they've all stopped? [15:01:27] qchris, any back of the napkin idea on how big webstats temp db files can get? [15:01:27] I don't know. We hardly use the bot. [15:01:36] order of magnitude guess is fine [15:01:38] 1G? [15:01:43] We used it on 2014-08-20, and yesterday. [15:02:02] The log from 2014-08-20 made it to the wikipage. Yesterday's didn't. [15:02:11] And the bot is not in the channel :-/ [15:02:27] they're different processes. checking... [15:02:44] ottomata: No clue. But I can provide a build that keeps the db around. That would allow us to check. [15:03:26] springle: Thanks! [15:04:37] qchris, i think lsof will tell me [15:04:40] looks like around 100MB [15:05:02] Ok. [15:05:03] not sure if that is correct though... [15:05:13] It was restarted 5 minutes ago. [15:05:27] So the database should be bigger at the end of the hour. [15:08:45] its 200M on gadolinium [15:09:43] I'll make a 2G tmpfs mount just to be safe [15:09:53] The unpacked projectcount files are a few hundred MBs. [15:10:02] 2G looks safe. [15:10:06] Agreed. [15:18:43] springle: Thanks for bringing analytics-logbot back! [15:20:26] !log yesterday (while the bot was not working), webstatscollector on gadolinium has been restarted twice (see http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-analytics/20140827.txt ) [15:32:05] qchris: mind if I restart webstats collector on gadolinium again? [15:32:09] to put it in a proper tmpdir? [15:32:26] Could we wait until 16:00? [15:32:29] sure [15:32:42] Cool. Thanks! [15:35:17] qchris: have you started doing comparisons of small project requests? i will do that if you haven't, and you can keep working on hive stuff... [15:35:51] Not yet. Cool. Thanks. [15:36:01] cool, did you have a project in mine, or shoudl I just pick one? [15:36:06] i want to pick one where udp2log has more requests than kafkatee [15:36:48] aa.b was a very small one. (<10 requests, and none on minutes immediately before/after the full hours) [15:36:59] But that might have changed since. [15:37:18] Generally, wikibooks domains seem to do the trick. [15:37:34] (Since they have more https traffic, it might be https related) [15:37:57] oh hmm [15:40:40] ok, i will look at finish wikibooks then :) [15:40:40] < fi.b - 724 6614284 [15:40:40] > fi.b - 713 6560067 [15:40:50] 11 more in udp2log [15:40:56] Cool! [15:59:12] ok, qchris, i'm running both a udp2log filter and a kafktee output capturing any logs matching fi.wikibooks.org on an03 [15:59:15] will let that run for an hour [15:59:29] then we will have raw logs from both to compare and pipe through filter and collector [15:59:39] Yay! Great! [15:59:59] So about restarting collector on gadolinium. [16:00:19] yes [16:00:19] Feel free to restart once the pagecounts and projectcounts are written. [16:00:22] ok [16:00:33] they are written! [16:00:44] restarting [16:00:52] !log restarted webstats-collector on gadolinium [16:01:34] cool, looks good [16:02:41] aw, ha, there have been a very few drops on the udp2log process.....but i guess that's normal, dunno! [16:02:49] on an03 [16:02:50] 823 [16:02:51] hmmm [16:03:41] oh i should up the recv queue on that [16:04:14] ah no, an03 still has the default way up [16:04:16] so, shoudl be ok [16:07:33] I looked for some time now, and the receive queue does not fill up. So looks good to me. [16:07:41] Maybe it was just choking during start? [16:09:01] its very rare [16:09:06] drops increase a little bit here and there [16:09:24] this is on (hex) port 20E4 [16:32:24] (CR) Milimetric: [C: 2] ":( disregard what I said - jquery is already global as you point out." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156453 (owner: Milimetric) [16:32:30] (CR) Milimetric: [V: 2] ":( disregard what I said - jquery is already global as you point out." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156453 (owner: Milimetric) [16:55:45] ottomata: No need to look further :-( [16:55:56] filter is counting some ssl requests twice [16:56:29] This explains why ssl heavy projects sometimes have higher numbers than the kafka pipeline. [16:57:06] oh, wha? [16:57:24] oh, it doesn't throw out the all the proxied varnish ones properly? [16:57:28] Analytics / General/Unknown: Webstatscollector's pagecounts file shaky between 2014-08-24 14:00 and 2014-08-27 21:00 - https://bugzilla.wikimedia.org/70118#c2 (Toby Negrin) Any ideas as to the root cause? Do you think this has been an ongoing problem that we haven't noticed? [16:57:55] (PS2) Milimetric: Add visualizer to coordinate selectors and graphs [analytics/dashiki] - https://gerrit.wikimedia.org/r/156722 [16:57:58] yes. [16:58:09] For example cp4008 is not thrown out. [16:58:16] Try running [16:58:42] grep 'commons.wikimedia.org/w/api.php.maxlag=3.format=xml.action=edit' /home/qchris/udp2log/cp4008.tsv [16:58:46] on stat1002. [17:01:27] huh, hm, i guess filter's dupes list needs updating, eh? [17:01:50] since ulsfo has been added since filter was last deployed [17:02:24] oh yeha, and in ulsfo, nginx proxies run on the same boxes as varnish [17:02:48] Yes. Exactly :-( [17:53:09] (CR) Nuria: "My only comment is to make the $.extend() explicit so we can use code as documentation and see parameters at a glance." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156346 (owner: Milimetric) [17:58:50] (CR) Milimetric: "Three main things:" (8 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/126927 (owner: Awight) [18:01:20] (PS3) Milimetric: Add visualizer to coordinate selectors and graphs [analytics/dashiki] - https://gerrit.wikimedia.org/r/156722 [18:04:15] (CR) Nuria: [C: 2 V: 2] "Correction: we will keep extend for visualizations as parameters might change a lot." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156346 (owner: Milimetric) [18:05:27] hm, welp, qchris, on an03 [18:05:37] hmm [18:05:38] More fun? [18:05:41] :-D [18:05:44] ah nm [18:05:55] was going to suggest a solution, but we'd have to modify webstats on gadolinium... [18:05:56] ah [18:05:56] hm [18:06:00] welp, i mean, we can do that... [18:06:03] we could add more filters [18:06:05] we probably should... [18:06:11] dunno though, we are trying not to change it [18:06:18] it does make a difference i think though [18:06:24] en was off by > 100K requests [18:06:26] more than 1 percent [18:06:26] Yes, totally. +1 for not changing things. [18:06:26] i think [18:06:28] for an hour [18:06:49] But I also think that ignoring requests with client ip 10.128.0 is worth it. [18:07:37] I stared at logs for the last hour, and ignoring 10.128.0. should be safe. [18:08:07] So what's up with analytics1003? [18:09:08] oh nothing, i was going to suggest something but then realized it didn't make any sense [18:09:17] probably caused by not yet having lunch [18:09:19] Ah. ok. [18:09:28] So let's have lunch. [18:09:38] And afterwards, we fix webstatscollector :-) [18:09:54] and now leila is all 'oh i really want to watch this meeting', and i have to wait for her because we are hanging out in nyc and I want to, even though she says I should go without her because i'm starving [18:09:55] but [18:09:56] NO [18:10:06] (PS2) Nuria: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 [18:10:22] owww ottomata. join us! ;p [18:10:40] So then help me understand the lvs traffic in ulsfo while you're waiting and we'ro straving together. [18:10:56] ok but first i'm going to have weekly checkin w gage :) [18:11:09] :-D [18:11:13] Ok. Have fun. [18:11:28] I'll prepare the patch then, and we can discuss there. [18:56:26] (PS1) QChris: Stop counting ulsfo's varnish relayed https requests [analytics/webstatscollector] - https://gerrit.wikimedia.org/r/156840 [18:58:41] (CR) QChris: Stop counting ulsfo's varnish relayed https requests (1 comment) [analytics/webstatscollector] - https://gerrit.wikimedia.org/r/156840 (owner: QChris) [18:58:45] (CR) Ottomata: [C: 2 V: 2] Stop counting ulsfo's varnish relayed https requests [analytics/webstatscollector] - https://gerrit.wikimedia.org/r/156840 (owner: QChris) [19:00:38] (CR) Ottomata: Stop counting ulsfo's varnish relayed https requests (1 comment) [analytics/webstatscollector] - https://gerrit.wikimedia.org/r/156840 (owner: QChris) [19:00:46] thanks qchris! [19:00:57] I do not like it. [19:01:08] I checked the logs so often, still I have a bad feeling. [19:01:18] Those filters just look wrong :-) [19:01:26] Thanks for merging. [19:01:44] Can I leave the packaging to you? [19:02:56] yup [19:02:58] i will package [19:03:01] and i know what you mean [19:03:06] in this case though, i think it will be ok [19:03:06] Thanks! [19:03:06] i mean [19:03:22] i dont' feel like that is a good way to correctly identify all internally proxied requests [19:03:24] but. [19:03:27] that is how it is done [19:03:33] and we are trying to change as little as possible :) [19:03:35] Hahahaha :-) [19:03:39] True. True. [19:03:55] here we are only doing something that (probably) diederik would have done if he were around when ulsfo was turned on [19:03:56] :) [19:05:10] I think this happened way after ulsfo saw traffic. [19:05:16] lunch finally! [19:05:44] Ulsfo saw traffic since 2013, right. But those requests only showed up in February 2014. [19:05:50] Enjoy your lunch! [19:11:40] (CR) Nuria: "I think the patch is missing corresponding gulpfile changes." [analytics/dashiki] - https://gerrit.wikimedia.org/r/156722 (owner: Milimetric) [19:17:52] (fixing the gulpfile now nuria, I'm gonna make it glob all the js files automatically) [19:31:47] ottomata: After you restarted the collector in the new layout, did you check that the dump files get generated? [19:32:03] Oh ... no ottomata around :-) [19:38:30] Analytics / General/Unknown: No webstatscollector output files after 2014-08-28 16:00 - https://bugzilla.wikimedia.org/70136 (christian) NEW p:Unprio s:normal a:None On http://dumps.wikimedia.org/other/pagecounts-raw/2014/2014-08/ the latest files showing up are for 2014-08-28 16:00. No... [19:53:26] qchris_away: do we have hiveserver2 running on any machines? [19:53:31] * YuviPanda wants to write some python scripts that hit hive [19:56:17] YuviPanda: analytics1027 has hiveserver2. [19:56:34] qchris: ah, cool. so I can connect from stat1002 to there... [19:56:52] Not sure. I assume not. [19:57:07] What do you want to achieve? [19:57:08] oh [19:57:20] qchris: write a python script that runs hive queries and does post processing on them? [19:57:21] (Typically the hive client would do) [19:57:29] hive client is interactive... [19:57:48] hive client takes files ... [19:57:56] still I've to shell out... [19:57:57] or HiveQL on the command line. [19:58:02] and can't relaly do 'fetchOne()' [19:58:07] (or fetchMany()) [19:58:52] Oh. I see. [19:59:00] Mhmm. [19:59:15] Could be that you can use stat1002 to do that kind of connection. [19:59:28] qchris: indeed, and I am able to start hiveserver2 on stat1002 [19:59:32] unsure if it actually works... [19:59:39] or if that is the appropriate solution even [19:59:50] Oh. Ok :-) [20:00:07] qchris: perhaps have puppet run hiveserver2 on stat1002... [20:00:26] ottomata: Did you see https://bugzilla.wikimedia.org/70136 ? [20:00:59] ehy don't start hiveserver2 on stat1002! [20:01:01] :) [20:01:06] can you do that? [20:01:08] why would you do that!? [20:01:26] uh oh, looking into it qchris... [20:02:44] Thanks. [20:03:11] ottomata: I just need some hiveserver to connect python to it to run queries :) [20:03:19] ottomata: and yes, running 'hiveserver2' from commandline starts one :) [20:03:30] hive-server is on analytics1027 [20:03:41] you can see the configs in /etc/hive/conf [20:03:53] ottomata: oh, and I can just connect to it? [20:03:58] also is it hiveserver or hiveserver2? [20:04:00] 2 [20:04:03] cool [20:04:08] hive CLI connects to it [20:04:09] yay then :) [20:04:11] don't see why python couldn't [20:04:12] aaaah [20:04:14] that makes senes [20:04:23] ottomata: it totally could, I just didn't know where it lived [20:05:06] hm, hive server 2 config is not there...i know hive CLI at least talks to metastore, but i thought it talked to hive-server2 [20:05:07] (PS4) Milimetric: Add visualizer to coordinate selectors and graphs [analytics/dashiki] - https://gerrit.wikimedia.org/r/156722 [20:06:42] ottomata: hmm, there's also 'embedded hive server' which might be what the commandline is using? [20:07:02] (PS3) Nuria: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 [20:07:42] qchris [20:07:42] http://utcc.utoronto.ca/~cks/space/blog/linux/Ubuntu1204Symlinks [20:07:52] pretty sure this is why collector isn't writing dumps [20:10:23] when we were running in shm, there was a subdirectory that was not world writeable [20:10:26] without sticky bit set [20:10:31] tmpfs i guess defaults to t hat [20:11:25] That would also explain why I got "permission denied" when I tried following the link, but it worked when directly accessing the link target. [20:12:51] (PS4) Nuria: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 [20:14:52] crap, qchris_away, i think we lost those hours of data [20:15:09] Yup :-) [20:15:14] i suppose we could fake regenerate them with kafka [20:15:19] since we can replay those logs [20:15:22] Hahahaha :-D [20:15:28] gettin messy! [20:16:00] Hehehe. And then we can use those regenerated logs to prove that the analytics1003 webstatscollector works fine :-D [20:16:44] hahahha [20:16:45] yeah! [20:16:47] So by reading the description (and assuming you're seeing that message in the logs), it seems it can easily be fixed. [20:16:47] great idea! [20:16:51] yes [20:17:33] That's great. [20:17:37] looking into unssetting sticky bit, but i think it would be safer to create a subdir with proper ownership inside of the tmpfs [20:17:49] mode= [20:17:50] hmmm [20:18:30] I have no preference really, and I know too little on WMF ops requirements :-/ [20:18:42] we can choose, actually, this mode option will probably be easier [20:18:44] i'm testing it now [20:21:04] ok, yeah, that should work [20:21:45] Works for my user too. [20:27:03] qchris: https://gerrit.wikimedia.org/r/#/c/156899/ [20:27:15] * qchris_away looks [20:28:45] (PS1) Yuvipanda: Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 [20:28:53] (CR) jenkins-bot: [V: -1] Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 (owner: Yuvipanda) [20:30:29] (PS1) BearND: Prepare for report: Number of blocked 'app users' on enwiki [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156916 [20:30:46] (PS5) Milimetric: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 (owner: Nuria) [20:33:28] Analytics / General/Unknown: No webstatscollector output files after 2014-08-28 16:00 - https://bugzilla.wikimedia.org/70136#c1 (christian) Potential fix by ottomata at https://gerrit.wikimedia.org/r/#/c/156899/1 [20:34:21] (PS2) Yuvipanda: Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 [20:34:28] (CR) jenkins-bot: [V: -1] Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 (owner: Yuvipanda) [20:34:48] (PS6) Milimetric: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 (owner: Nuria) [20:35:08] (PS3) Yuvipanda: Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 [20:38:38] nice YuviPanda ^ proper session closing was something we did way too late in wikimetrics [20:39:03] milimetric: yeah, has been biting me a bit as well [20:39:32] I'm realizing more and more how bad I am at back-endy stuff. In .net world everything was made easy for me [20:39:40] * milimetric should stick to JS [20:39:40] :) [20:40:01] milimetric: :D [20:40:17] milimetric: how are you writing CSV in wikimetrics, btw? are you streaming it or just writing it all in memory and writing it out in one go? [20:40:24] I've a hack to make it stream while still using the csv module... [20:40:33] milimetric: or we could write things in Mono and still use C# ;) [20:40:42] YuviPanda: just in memory [20:40:52] then dumping it in one go [20:41:09] (PS2) BearND: Prepare for report: Number of blocked 'app users' on enwiki [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156916 [20:41:15] milimetric: aaah, neat and quick hack that lets us stream it https://github.com/wikimedia/analytics-quarry-web/blob/master/quarry/web/output.py#L15 [20:41:31] needs comments tho [20:42:49] cool, i have no idea how that works :) [20:43:00] milimetric: duck typing! :D [20:43:16] I construct an object that has .write() which is also all that the csv module wants [20:43:23] so it writes out a line, I read out a line and yield, repeat.... [20:43:55] (PS7) Milimetric: [WIP] Project and language retrieval from wiimetrics 'api' [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 (owner: Nuria) [20:44:37] ah, cool [20:45:43] milimetric: YuviPanda: spare some time for some CR in https://gerrit.wikimedia.org/r/#/q/status:open+project:analytics/limn-mobile-data,n,z? [20:46:07] YuviPanda: and you've seen this: http://flask.pocoo.org/docs/0.10/patterns/streaming/ [20:46:09] * YuviPanda passes that baton to milimetric [20:46:17] heh, i'll take a look bearND [20:46:26] milimetric: I have! problem is the CSV module doesn't support streaming, and flask expects to yield [20:46:37] only need the first and the third patch, the 2nd one can come later once the csv's are generated [20:47:40] oh I see you're hacking the csv module [20:47:42] makes sense [20:47:44] milimetric: yes [20:47:58] (CR) Yuvipanda: [C: 2] Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 (owner: Yuvipanda) [20:48:08] (Merged) jenkins-bot: Attempt to ensure that sessions are closed properly [analytics/quarry/web] - https://gerrit.wikimedia.org/r/156909 (owner: Yuvipanda) [20:50:40] (CR) Milimetric: [C: -1] "small sql thing" (1 comment) [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156735 (owner: BearND) [20:51:08] qchris_away: built a new package... [20:51:13] shoudl I add it to apt and upgrade..hmm [20:51:16] it is 5pm here... [20:51:23] maybe tomorrow morn? :) [20:51:28] although, tomorrow is firday [20:51:31] and i'm only working a half day [20:51:31] haha [20:51:34] maybe we shoudl wait.. [20:53:11] (CR) Milimetric: [C: 2] Prepare for report: Number of blocked 'app users' on enwiki [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156916 (owner: BearND) [20:53:57] milimetric: thank you. I'm going to rebase #1 and amend it shortly thereafter. [20:54:39] (CR) Milimetric: [C: 1] "you can self-merge this after you look at the suggestion" (1 comment) [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156736 (owner: BearND) [20:55:42] ottomata: add it :-) [20:55:52] haha uhhh [20:55:55] you sure? [20:56:00] No. Let's wait. [20:56:03] It's ok. [20:56:13] Data has been bad for 6 months. [20:56:21] A few more days won't kill us. [20:56:27] haha [20:57:10] Now back to eating :-) Nom Nom Nom [20:59:42] :) [21:00:14] (PS2) BearND: Add edit funnel sql for apps [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156735 [21:00:30] (PS3) BearND: Add edit funnel sql for apps [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156735 [21:00:32] (PS2) BearND: Add edit funnel reports for apps to dashboards [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156736 [21:03:10] milimetric: added the parens in https://gerrit.wikimedia.org/r/#/c/156735/. Did you mean I could self-merge https://gerrit.wikimedia.org/r/#/c/156735/ as well? I didn't see any comments in https://gerrit.wikimedia.org/r/156736. [21:12:40] (CR) Milimetric: [WIP] Project and language retrieval from wiimetrics 'api' (2 comments) [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 (owner: Nuria) [21:15:09] milimetric: so, I guess you want me to self-merge both outstanding patches, right? [21:15:41] yeah, they look good to me bearND, that repo is pretty low-risk [21:15:51] milimetric: ok, great [21:16:17] if something is horribly broken, just let me know and I'll help fix [21:16:56] milimetric: sounds good! Thanks for catching the missing parenthesis earlier [21:17:02] np [21:17:09] (CR) BearND: [C: 2] Add edit funnel sql for apps [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156735 (owner: BearND) [21:17:14] (Merged) jenkins-bot: Add edit funnel sql for apps [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/156735 (owner: BearND) [21:17:25] bearND: if i remember correctly, some of your other sql had that error, but i didn't catch it at first [21:17:34] you might want to take a look at the ones that report ios numbers [21:17:52] milimetric: will check [21:19:30] ok ^d, looking good! [21:19:34] oops, wrong chat [21:19:50] milimetric: no, that was the only one in a Where clause. The other ones for iOS were in the Select clause [21:20:19] cool [21:31:18] ttyall tomorrow! [21:31:37] Good night! [21:43:14] (CR) Milimetric: [WIP] Project and language retrieval from wiimetrics 'api' (9 comments) [analytics/dashiki] - https://gerrit.wikimedia.org/r/156741 (owner: Nuria) [21:59:15] Analytics / General/Unknown: Webstatscollector counting HTTPS from ulsfo twice - https://bugzilla.wikimedia.org/70140 (christian) NEW p:Unprio s:normal a:None HTTPS requests end up twice in the udp2log pipeline for all data-centers. Once from directly from the nginx-s, once from the varnish... [21:59:28] Analytics / General/Unknown: Webstatscollector counting HTTPS from ulsfo twice - https://bugzilla.wikimedia.org/70140#c1 (christian) A corresponding fix at https://gerrit.wikimedia.org/r/#/c/156840/1 has been merged, but due to the pending weekend, and since Monday is Labour Day, it'll probably tak... [22:22:14] Analytics / General/Unknown: Webstatscollector's pagecounts file shaky between 2014-08-24 14:00 and 2014-08-27 21:00 - https://bugzilla.wikimedia.org/70118#c3 (christian) (In reply to Toby Negrin from comment #2) > Any ideas as to the root cause? Yes. Bug 70053. Due to the increasing number of process... [22:26:14] Analytics / General/Unknown: No usable webstatscollector output files for 2014-08-28 17:00 -- 21:00 - https://bugzilla.wikimedia.org/70136#c2 (christian) NEW>RESO/FIX The first new files for a full hour are in again: http://dumps.wikimedia.org/other/pagecounts-raw/2014/2014-08/pagecounts-20140...