[00:03:00] (CR) Nuria: Bootstrapping from url. Keeping state. (1 comment) [analytics/dashiki] - https://gerrit.wikimedia.org/r/160685 (https://bugzilla.wikimedia.org/70887) (owner: Nuria) [00:13:05] (PS13) Nuria: Bootstrapping from url. Keeping state. [analytics/dashiki] - https://gerrit.wikimedia.org/r/160685 (https://bugzilla.wikimedia.org/70887) [00:14:18] kevinator: i totally forgot about deployment today too [00:14:37] I pinged dan earlier today [00:14:39] he did it [00:14:51] I was about to write an email to wikimetrics about it. [00:30:13] man, hadoop is FAST [00:30:34] SELECT project, count(*), group by project over a month of mobile and desktop requests, started an hour ago. Quarter of the way through mapping. [00:30:47] * Ironholds shakes head in amazement. [00:37:02] Analytics / Wikimetrics: Story: WikimetricsUser has better explanation counting on all namespaces - https://bugzilla.wikimedia.org/71582 (Kevin Leduc) NEW p:Unprio s:minor a:None This applies to edits and pages created metrics. You need to delete the text in the "namespaces" textfield to se... [00:38:00] Analytics / Wikimetrics: Story: WikimetricsUser has better explanation counting on all namespaces - https://bugzilla.wikimedia.org/71582#c1 (Kevin Leduc) Created attachment 16654 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16654&action=edit pic of bug [01:44:01] (PS1) Yuvipanda: Update fab to send to new hosts [analytics/quarry/web] - https://gerrit.wikimedia.org/r/164501 [01:44:17] (CR) Yuvipanda: [C: 2] Update fab to send to new hosts [analytics/quarry/web] - https://gerrit.wikimedia.org/r/164501 (owner: Yuvipanda) [01:44:22] (Merged) jenkins-bot: Update fab to send to new hosts [analytics/quarry/web] - https://gerrit.wikimedia.org/r/164501 (owner: Yuvipanda) [09:12:23] (Abandoned) Milimetric: Improve survival metric query performance [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/121630 (https://bugzilla.wikimedia.org/63201) (owner: Milimetric) [09:13:46] (PS2) Milimetric: Cleaning up before doing anything in the build [analytics/dashiki] - https://gerrit.wikimedia.org/r/164254 (owner: Nuria) [09:14:00] (CR) Milimetric: [C: 2 V: 2] Cleaning up before doing anything in the build [analytics/dashiki] - https://gerrit.wikimedia.org/r/164254 (owner: Nuria) [09:15:47] (PS14) Milimetric: Bootstrapping from url. Keeping state. [analytics/dashiki] - https://gerrit.wikimedia.org/r/160685 (https://bugzilla.wikimedia.org/70887) (owner: Nuria) [09:15:56] (CR) Milimetric: [C: 2 V: 2] Bootstrapping from url. Keeping state. [analytics/dashiki] - https://gerrit.wikimedia.org/r/160685 (https://bugzilla.wikimedia.org/70887) (owner: Nuria) [09:16:05] (CR) Milimetric: Bootstrapping from url. Keeping state. (1 comment) [analytics/dashiki] - https://gerrit.wikimedia.org/r/160685 (https://bugzilla.wikimedia.org/70887) (owner: Nuria) [09:39:32] Analytics / Wikimetrics: slow report result and report list fetching - https://bugzilla.wikimedia.org/71603 (Dan Andreescu) NEW p:Unprio s:normal a:None It looks like we need a simple index on the report table, probably on the user_id column, to speed up two operations that have become slow... [11:30:02] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606 (christian) NEW p:Unprio s:normal a:None Dan reported through email that we're seeing [Sun Sep 28 14:24:23 2014] [error]... [11:47:45] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606#c1 (christian) The first occurrence I could find in the logs has a timestamp of [Sat Sep 20 14:37:58 2014] ... almost two weeks bac... [11:51:00] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606#c2 (christian) The webserver's config around api.wsgi is from 2014-01-20, so that should not have changed in the last two weeks. [12:00:55] mornin qchris! [12:01:02] Heya! [12:01:45] ok, so i think when i do this naming thing, I have an OO / REST background in mine [12:01:54] which is why I like noun/verb so much [12:02:38] well, maybe REST isn't really relevant here [12:02:45] but I llike organizing from largest to highest concept [12:02:52] and nouns seem to be a larger concept to me [12:03:02] Ah. naming :-) [12:03:03] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606#c3 (christian) D'oh! The IP is not the machine's IP, but that of dynamicproxy-gateway.eqiad.wmflabs. [12:03:09] hahaa [12:03:10] (PS1) Milimetric: Style layout to fill page [analytics/dashiki] - https://gerrit.wikimedia.org/r/164536 [12:03:12] yeah i just read your review [12:03:12] (PS1) Milimetric: Style color swatches square and borderless [analytics/dashiki] - https://gerrit.wikimedia.org/r/164537 [12:03:14] (PS1) Milimetric: Style metric selector button and popover [analytics/dashiki] - https://gerrit.wikimedia.org/r/164538 [12:03:16] (PS1) Milimetric: Style border radii and project list [analytics/dashiki] - https://gerrit.wikimedia.org/r/164539 [12:03:18] hm, but you want to have big_noun/verb_yeah_yeah [12:03:51] i want big_noun/smaller_noun/even_smaller_noun_if_needed/verb [12:03:56] or not! [12:04:21] why you gotta be hatin on adjectives yo [12:04:24] i think I don't want to make a blanket policy across everything. it will be tough to define something that fits every circumstance [12:04:29] sure, adjectives! [12:04:41] :-) [12:04:50] Ok. Let's focus on /oozie then. [12:04:56] hha [12:04:59] well, i mean even within /oozie [12:05:15] Ok. [12:05:31] I am all for structure. [12:05:44] But having two abstraction levels for only [12:05:59] two oozie jobs, just seem overstructured. [12:06:04] btw, i'm about to add a drop workflow [12:06:12] util/hive/partition/drop [12:06:18] Then s/two/three/ [12:06:28] util is separate for me. [12:06:31] that i want to use as a subworkflow for sqoop stuff [12:06:32] yes agree. [12:06:38] hive/partition/drop [12:06:51] webrequest/partition/add [12:07:10] hm, you say that you would like that to be check_integrity? [12:07:20] A bit. yes. [12:07:22] but, the reason that dir exists is to add a partitoin. [12:07:34] Originally, yes. [12:07:39] it just happens that we check and report on the partition's integrity too [12:07:43] But now the key part (at least from my point of view) [12:07:49] is that it checks integrity. [12:08:10] So ok. Partition adding is more important to you. [12:08:11] Mhmm. [12:08:22] Was only a side effect for me. [12:08:46] that is where more of the work is in that workflow, but that is almost more a monitoring thing [12:09:06] (PS2) Milimetric: Style border radii and project list [analytics/dashiki] - https://gerrit.wikimedia.org/r/164539 [12:09:19] like, i wouldn't name a the kafka puppet module after the monitoring class that it has in it [12:09:35] Unfair comparison :-D [12:09:46] But I agree. The oozie workflow does two things. [12:09:53] (agree unfair :p) [12:09:59] We have different views on which of the two is more important. [12:10:22] well, i guess, ok. I think OO is going out of style, but that is where I draw most of my influences :p [12:10:28] object.verb() [12:10:38] Ok. Let's ignore the names for a second. [12:10:47] Do you think we need that much layers of structure? [12:10:54] (Regardless of whether noun or verb) [12:11:24] I mean we totally could, but it's just that [12:11:35] ok, so, for example, you would prefer to have [12:11:39] it makes it hard for me to cd down the whole hierarchy when trying to find things. [12:11:43] hive/add_partition [12:11:43] hive/drop_partition [12:11:44] ah [12:11:54] you know, qchris, i think for me it is more concept ordering than it is hierarchy [12:11:58] so [12:11:59] in that example [12:12:21] i wouldn't mind the same hierarchy, but things named [12:12:21] hive/partition_drop [12:12:21] hive/partition_add [12:12:37] that way concepts are still grouped together [12:12:40] You're taking about util :-) [12:12:45] yes [12:12:46] util is separate for me. [12:12:48] oh ok' [12:12:52] separate from naming altogether [12:12:55] ok, well that was just an example [12:13:01] I was more thinking of webrequest directory. [12:13:03] same for non-util [12:13:15] i would be ok with webrequest/partition_add [12:13:18] and even [12:13:24] webrequest/done_flag_monitor :p [12:13:35] Hahaha :-) [12:13:38] hmm [12:13:41] ah hm [12:13:57] Ok. I see. You really, really care about the ordering. [12:14:15] If we have the ordering, we can also have the "/". [12:14:18] yeah i think so, espescially in the directory names, but probably a little also in the files [12:14:41] as for the generate_tsvs thing [12:15:09] So we should change [12:15:13] webstats/generate_hourly_files [12:15:17] webstats/insert_hourly_pagecounts [12:15:19] to [12:15:22] webstats/hourly_files/generate [12:15:26] webstats/hourly_files/insert [12:15:37] (or something the like) too? [12:15:50] ya something like that (not so sure about 'hourly_files') but yeah. [12:16:06] although, in that particular example, the object is actually different there [12:16:14] webstats/hourly_files [12:16:15] webstats/table :/ [12:16:27] (I'm not suggesting that ^!) [12:16:27] My "overengineered" alarms get off pretty heavily :-) [12:16:31] But I guess you win here. [12:16:52] Next to those alarms, I have no real counter-argument. [12:17:22] wait, how'd I win? the ordering argument? [12:17:36] You wanting it really hard made you win. [12:17:39] hahahha [12:17:41] I lack hard counter-arguments. [12:17:44] awww come onnnn! [12:18:01] No really. I could only block. But with no real hard reason. [12:18:05] That'd be pointless. [12:18:16] qchris, btw, i think it is going to be really hard to obey this or any rule for all circumstances [12:18:36] True. But it should serve as guiding principle. [12:18:46] i like to organize things this way, and it has worked pretty well in the past, but only if you just try to stick to the convention...and yeah, just let it guide but not rule you [12:18:49] hm [12:18:54] well, what to do about tsvs? [12:19:21] Since it's hard to copy paste again and again in IRC ... [12:19:30] Let's work it out in [12:19:33] http://etherpad.wikimedia.org/p/analytics-bikeshed [12:19:35] and talk here [12:19:36] haha ok [12:20:41] Which verb to take under /webstats? [12:21:46] (PS1) Milimetric: Fix knockout for firefox [analytics/dashiki] - https://gerrit.wikimedia.org/r/164541 [12:21:58] awww, i agree we should call these pagecounts-all-sites, but webstats is so much cleaner :/ :) [12:22:00] too bad. [12:22:22] cleaner in what sense? [12:22:38] I mean ... we could keep webstats ... [12:23:12] the only reason we are choosing that name is to stick with the public convention, i think webstats is a better name aesthetically. But! I need to remember that this is a 'stop-gap' anyway [12:23:15] i shoudln't worry about it [12:23:18] anyway. thinking. [12:23:30] ok. [12:23:31] Can we somehow get the integrity check in the name of "./webrequest/partition/add"? [12:23:43] s/in/into/ [12:24:17] hashar: Hi! :-) [12:24:45] just kidding :D [12:24:59] You've got me at UUID :-) [12:25:12] watching our VPE introduction https://www.youtube.com/watch?v=GJGC9zpbJpU&feature=youtu.be :D [12:27:43] Seeing the suggestions ... I guess I do not like the "partition" in the name. [12:27:54] All of the jobs operate on/for partitions. [12:28:00] But only two have it in the name. [12:28:05] That is skew. [12:29:55] ottomata: Why would you not use a second level for pagecounts-all-sites? [12:30:14] Like something between pagecounts-all-sites and archive. [12:30:23] still thinking, reading your patch [12:31:02] k [12:34:25] qchris, stil thinking, but i am coming around to your no 'partition' argument for webrequest (not for util) [12:35:02] AGH [12:35:05] why is etherpad so bad at pasting [12:35:10] So to have an example: [12:35:17] it always converts my newlines to spaces [12:35:43] A pagecounts' file also exacly covers a partition. If we use "partition" in webrequest, shouldn't we also have [12:35:54] but i think i would say I'm not putting anohter level in, because the final level for both things there are operating on pagecounts-all-sites [12:35:55] pagecounts-all-sites/partition/dump [12:35:56] like [12:36:50] pagecounts_all_sites = PagecountsAllSites.new() [12:36:51] pagecounts_all_sites.generate(year-month-day-hour) [12:36:51] pagecounts_all_sites.archive(year-month-day-hour) [12:36:56] Wait ... you're current suggestion strips a level of abstraction. I love that! [12:37:08] yes, i am saying i am coming around [12:37:30] not liking webrequest/tsvs/ yet, but maybe... [12:37:52] You also argued in the CR that "tsvs" is bad. [12:37:59] What about legacy_tsvs [12:38:02] udp2log_tsvs [12:38:12] That makes it clearer what kind of tsvs it is about. [12:38:33] oh, yes i haven't even begun thinking about the actual noun [12:38:35] haha [12:38:43] i'm still just thinking hierarchy [12:38:48] Ok. [12:38:59] I like your hierarchy. [12:39:11] The hierarchy part, I buy it as is. [12:39:20] hm [12:39:58] ok, hm, taking a step back from it, thinking about 'tsvs' [12:40:11] legacy_tsvs [12:40:11] hm [12:40:24] sure. [12:41:18] in general, i'm reluctant to name the dataset after the format [12:41:29] but, in this case i think its ok, mainly because it is legacy [12:41:31] hm [12:41:38] Yup. I feel the same way. [12:41:56] legacy_dump or something like that would work for me too. [12:42:08] But everyone is calling them the tsvs. [12:42:32] qchris: as for the hour, daily, etc. suggestion, i doubt we would organize the hierarchy that way, as wouldn't those concepts be abstracted in the oozie configs somehow? [12:42:45] similar to how you are abstracting the actual tsv generation by selecting a specific hive script [12:42:46] ? [12:43:00] yeah, since everyone is calling them tsvs [12:43:01] I am not sure. Maybe. [12:43:02] let's just call them that [12:43:17] yeah, im' not sure either, meh, let's worry about that if/when it happens [12:43:33] pagecounts-all-sites is hourly by definition right now [12:43:36] Ok. I removed them. [12:44:03] qchris, how do we feel about underscore vs. hyphens here [12:44:11] pagecounts-all-sites vs legacy_tsvs [12:44:12] ? [12:44:36] On the dumps.wikimedia.org server, it's hyphens. [12:44:39] yes [12:44:45] So I figured we should use them here too. [12:44:58] just for that one dataset for historical consistency reason? [12:45:00] reasons* [12:45:00] ? [12:45:05] And IIRC you said you prefer underscore as separators. [12:45:08] i do. [12:45:28] Yes, "just for that one dataset for historical consistency reasons". [12:45:31] ok [12:45:36] ok ok let's leave that then [12:45:36] ok [12:45:51] Meh. But I am not sold on that. [12:45:52] haha, you like the structure only because there are fewer directories! [12:45:55] Underscores are fine too. [12:46:05] Yes. Fewer directories rock :-D [12:46:10] qchris: i'd be ok if the 'display name' used hyphens [12:46:23] but we used underscores everywhere else [12:46:40] * qchris does not understand 'display name' [12:46:47] You mean the directory name? [12:46:49] e.g. in dumps.wm.o [12:46:53] or name of the workflows. [12:46:54] in a 'UI' [12:46:55] oh. [12:46:59] ok. [12:47:02] if you want. :) [12:47:26] anyway, what about those names don't you like? [12:47:38] i kinda like ingest. [12:47:51] Ok. Me too. [12:49:47] i don't like ingest for pagecounts-all-sites [12:49:51] HMMM [12:49:53] ergh. hm [12:49:55] hm [12:49:59] :-) [12:50:21] Because it duplicates the data? [12:50:26] ok, i kinda like ingest for webrequest, because for that case, it is making data available for the first time [12:50:42] that is not entirely true, as the real 'ingestion' is being handled by camus [12:50:59] Oh. But note the structure! It's not general ingestion. But ingestion in the scope of "pagecounts-all-sites". [12:51:06] but, from the point of view of hive (and that is our main viewpoint in hadoop right now), the data is not available at all until it has a partition added [12:51:13] haha [12:51:19] (In that scope) the data becomes availbale for the first time too. [12:51:22] haha [12:51:27] EEEEeeeeEEEEEE that is true. [12:51:57] Maybe .... "load" like the L in ETL? [12:51:58] ingest just has a different meaning to me though, i think. ingest, to me, implies that something was sucked in from a different system [12:52:16] hm [12:52:17] load. [12:52:25] Even that would match here. The different system is the "webrequest" table. [12:52:43] Meh. I actually prefer load. [12:52:46] Load is great. [12:53:01] nawwwwwww system is a pretty general term, but i think wouldn't call webrequest a different 'system' [12:53:03] but, yeah [12:53:04] i think i like load [12:53:11] i'm ok with load for webrequest too. [12:53:29] i think i like archive better than dump though [12:53:57] Mhmm ... does archive describe what is happening? [12:54:03] Well. sort of. [12:54:15] But we could archive in so many different ways! [12:54:25] true, we could dump in many different ways too! [12:54:29] are you guys dropping Limn ? [12:54:31] Right. [12:54:39] ^ was for ottomata. [12:54:42] :D [12:55:00] qchris, i agree, hm, what if we had /wmf/data/dumps [12:55:02] I guess milimetric would know whether Limn is being dropped/replaced :] [12:55:04] instead of /wmf/data/archive [12:55:06] is that better? [12:55:08] not sure. [12:55:45] * qchris thinks. [12:56:31] From the table point of view, "archive" is actually good. Because the table will have it's data pruned at some point in time. [12:56:52] From the action point of view, dump seems more appropriate. [12:57:04] From the people point of view "make public" seems more appropriate. [12:57:31] Archive somewhat feels �old, dusty, and boring" to me. [12:59:11] ha [12:59:25] archive is such a great word! 'legacy' feels older and dustier to me :p [12:59:34] Meh. I am ok with archive. [12:59:48] Let's stick with archive for now :-) [13:00:04] haha, both archive and 'dump' do this: they have the convenience of being both a verb and a noun [13:00:15] so we could each think of them as we like and be happy :p [13:00:21] Should we also use "Archive" for webrequest instead of lecagy_tsvs? [13:00:30] hm [13:00:33] no [13:00:39] Yay. Being both noun and verb is nice. [13:00:48] Same for load :-) [13:01:10] because the legacy_tsvs are a different thing than webrequest is [13:01:17] we aren't just archiving webrequest [13:01:40] The same holds true for pagecoutns-all-sites. It does "per wiki" roll-ups. [13:01:51] (PS1) Gilles: Update actions schema version and add new actions [analytics/multimedia] - https://gerrit.wikimedia.org/r/164545 [13:02:16] (CR) Gilles: [C: 2] "Tested by tunneling to the SQL server" [analytics/multimedia] - https://gerrit.wikimedia.org/r/164545 (owner: Gilles) [13:02:21] (Merged) jenkins-bot: Update actions schema version and add new actions [analytics/multimedia] - https://gerrit.wikimedia.org/r/164545 (owner: Gilles) [13:04:09] bwer? [13:04:11] oh [13:04:13] the project counts [13:04:15] hah [13:04:15] hm [13:04:27] right. [13:04:45] diifficult chooiiiiices [13:04:57] Hahaha :-) [13:05:03] webrequest/legacy_tsvs/archive? [13:05:03] haha [13:05:05] oof [13:05:09] ahhhHHHGHHG [13:05:20] Meh. I am fine with having legacy_tsvs for webrequest [13:05:25] well, it just so happens that you are archiving AND generating in the same workflow [13:05:26] and archive for pagecounts-all-sites. [13:05:35] i'm fine with it as we have it too, qchris. [13:05:49] Ok. [13:05:53] So then it is: [13:06:00] ./webrequest/load [13:06:00] ./webrequest/monitor_done_flag [13:06:00] ./webrequest/legacy_tsvs [13:06:00] ./pagecounts-all-sites/load [13:06:00] ./pagecounts-all-sites/archive [13:06:01] ? [13:07:08] Oh no. Sorry. That does not match what you suggested in the etherpad. [13:08:10] nono [13:08:12] your suggestion [13:08:14] i haven't changed mine [13:08:16] that's cool [13:08:23] there. [13:08:27] i changed mine to match yours [13:08:36] Ok. Cool. [13:08:48] I'll prepare the patches for this change over the weekend. [13:08:53] what a productive naming discussion! [13:08:57] didn't even need any UUIDs! [13:09:03] I liked it! [13:09:04] Thanks. [13:09:21] But ... UUIDs ... we could totally do that. It would be useful! [13:10:47] (CR) QChris: [C: -1] "I will amend to make the directory" [analytics/refinery] - https://gerrit.wikimedia.org/r/162589 (owner: QChris) [13:11:35] (PS1) Gilles: Add new actions for download popup [analytics/multimedia/config] - https://gerrit.wikimedia.org/r/164546 [13:11:37] Now hashar left ... we should have use UUIDs :-( [13:12:16] (CR) Gilles: [C: 2] "Tested with local limn pointing to production tsvs" [analytics/multimedia/config] - https://gerrit.wikimedia.org/r/164546 (owner: Gilles) [13:12:41] (CR) Gilles: [V: 2] Add new actions for download popup [analytics/multimedia/config] - https://gerrit.wikimedia.org/r/164546 (owner: Gilles) [14:00:06] ottomata / nuria: I'm running standup "quickly" today [14:00:17] k [14:00:19] there's a fun saiku meeting happening right now [14:00:29] (which you're all invited to if you're interested) [14:00:56] iam in the standup ... let's see if it clicks in [14:01:39] hashar: no, Limn will be around for a minute [14:01:44] (a long time) [14:02:33] milimetric: thx.I was just being curious. [14:29:02] today i learned Kevin is your product manager. I though he was a second ops :D [14:30:22] hashar, yes, I got a lot of commentary during his metrics presentation [14:30:23] along the lines of [14:30:36] "Who is this guy? Product manager for analytics? HE'S NOT ON HOWIE'S TEAM! HE ESCAPED!" [14:30:50] I had to point out we had several PdMs who are attached directly to their teams [14:36:54] Ironholds: sometime that makes more sense :] [14:37:02] agreed! [14:39:09] I am watching the metrics presentation rightnow [14:39:33] sounds like the metrics presentation was my coming out party [14:41:30] kevinator: nice pres :-] [14:41:37] I love how the web interface is super simple [14:41:45] thank you [14:41:59] yet having a huge complicated backend being it "just" to draw a couple lines hehe [14:42:16] Toby and I looked at each other when Erik announced he’d be using Vital Signs at the next metrics meeting [14:42:57] backend… yeah… we’re exploring data warehouses right now to improve performance [14:43:00] isn't it going to be presented by Damon ? [14:43:43] I don’t know… Erik may hang on to the responsibility of presenting the metrics on the community [14:44:11] and milimetric Graph extension showing up ( https://www.mediawiki.org/wiki/Extension:Graph ) [14:44:19] I should watch those metric meetings more often [14:45:56] and I am off, will be back later [15:04:54] tnegrin: yoyo [15:05:10] 5 mins -- just finishing up [15:06:02] k [16:01:02] (CR) Nuria: [C: 2] Style border radii and project list [analytics/dashiki] - https://gerrit.wikimedia.org/r/164539 (owner: Milimetric) [16:01:11] (CR) Nuria: [V: 2] Style border radii and project list [analytics/dashiki] - https://gerrit.wikimedia.org/r/164539 (owner: Milimetric) [16:01:25] (CR) Nuria: [C: 2 V: 2] Style metric selector button and popover [analytics/dashiki] - https://gerrit.wikimedia.org/r/164538 (owner: Milimetric) [16:01:43] (CR) Nuria: [C: 2 V: 2] Style color swatches square and borderless [analytics/dashiki] - https://gerrit.wikimedia.org/r/164537 (owner: Milimetric) [16:02:02] (CR) Nuria: [C: 2 V: 2] Style layout to fill page [analytics/dashiki] - https://gerrit.wikimedia.org/r/164536 (owner: Milimetric) [16:05:35] * Ironholds pokes ottomata [16:06:04] See the thread "Fwd: Monthly MaxMind Receipt" when you're back? [16:06:35] Ironholds: oh yeah, i'm in 1:1 w toby, wil lask him about that now [16:07:58] we are talking about this now [16:08:00] Ironholds: [16:08:09] hah! cool :) [16:08:11] thanks both [16:09:27] so this is just connection type? [16:09:40] Ironholds: join our little meeting! [16:09:40] https://plus.google.com/hangouts/_/wikimedia.org/otto-tnegrin [16:09:46] Ironholds: come hither: https://plus.google.com/hangouts/_/wikimedia.org/otto-tnegrin?authuser=0 [16:10:13] kk [16:10:15] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606#c4 (christian) Seems the issue is triggered by pointing a browsing to https://metrics.wmflabs.org/server-status [16:12:30] Analytics / Wikimetrics: Apache's logs containing "client denied by server configuration: /srv/wikimetrics/wikimetrics/api.wsgi" - https://bugzilla.wikimedia.org/71606#c5 (nuria) Then is a lawful error, I think there is no harm on that one returning a 403. Loads of interesting things on server-status:... [16:14:34] * Ironholds dances [16:14:46] now I just need the table locks on enwiki and kowiki dealt with and we are /cooking with fire/ [16:14:53] or at least, some sort of high-temperature concept or substance [16:48:25] Ironholds: did you talk to toby about what level to purchas? [16:48:29]