[17:02:46] joakino, phuedx, dr0ptp4kt: where'd you guys go? D: [17:05:11] jdlrobson: pingerino [17:55:46] in https://phabricator.wikimedia.org/T123077 the task says to remove Class and Class.extend. What is the new inheritance model to extend functionality? [17:56:36] is it now OO.mfExtend? [18:19:28] nvm, got it :) [19:22:05] bearND: my only hesitation about further rolling out RB usage is the fact that the Main Page seems to get stuck (because of cache?). Will that be addressed? [19:22:39] dbrant: what do you mean it gets stuck? [19:23:44] bearND: it's stale -- the app currently shows "Andrew Johnston", which was the featured article on Dec 26. [19:24:34] dbrant: ok, i'll look into this. We'll postpone further roll out until we have resolved this. [19:48:15] bearND|afk dbrant: i'm seeing an issue with the menu bar being way too short in landscape orientation on API 22 devices. i'm looking into it [20:01:46] brb [20:06:38] niedzielski: i see what you mean. i'm guessing it has to do with the "maximum" height imposed on the ArticleHeaderView [20:07:17] dbrant: yeah, i was thinking it might have something to do with the status bar logic but now i'm not sure [20:37:01] dbrant: good call https://gerrit.wikimedia.org/r/263412 [21:06:13] bearND: your Friday 11-1245 (PST) meeting, is that a new conflict for Friday retros? [21:08:14] yo [21:08:25] bmansurov, jhobs: yt? [21:08:33] yup [21:08:46] what up? [21:09:49] mbinder: it's not a new one. I'm usually ok 11:30-12:45 PST. [21:10:35] bearND: OK, just figuring out long term scheduling, I'll keep that in mind, thanks [21:10:52] bearND dbrant: release notes are cooking in the usual place [21:14:18] phuedx: ? [21:14:29] just wandering what you're working on? [21:15:31] right now I'm reading some of king's code and demos online about node.js to better familiarize myself with it (one of my focus areas for this quarter) [21:15:45] needed a bit of a cleanser after the long meeting ;) [21:15:50] i'll be doing some code review later [21:16:01] hbu? [21:17:44] yeah [21:18:17] phuedx: i've followed up on my patches, taking a break until later today ; [21:18:52] jhobs: not sure -- i'd like to write some phabricator tasks for what i think we're supposed to be working on [21:19:05] and maybe write up some notes on the cite extension [21:30:52] niedzielski: looks mostly good to me. There is a minor comment on the last line. Not sure if that was intentional. :) [21:31:54] bearND: :) [21:49:08] niedzielski: made a couple notes [21:50:35] dbrant:ok cool [21:55:23] bearND|brb dbrant: cool, beta is in progress again so please hold off on merges for a short bit. [21:57:44] joakino: I can't get your stack thing to run locally! It's driving me nuts :D [21:58:03] node keeps having a problem with the "=>" syntax whenever i run npm start [22:09:43] jhobs: you should update to nodejs 4.2.1 [22:12:05] bearND: do we have some kind of versioning miscommunication, or am I _incredibly_ outdated? My current version is 0.10.32... [22:13:24] wow no I was just that outdated. I'm on 5.4.0 now [22:14:32] and that did the trick; thanks, bearND! [22:15:08] jhobs: are you on ubuntu? there are other issues the debian package, i think the executable is called nodejs instead of "node". maybe it's no longer maintained? [22:15:35] niedzielski: nah, mac [22:15:38] The => is ES2015/aka Harmony syntax. 0.10 is quite old. Some of the servers were updated to 4.2 (LTS) last week BTW. [22:16:09] jhobs: did you use brew to install node? [22:16:37] probably initially [22:16:49] I updated it just now with npm itself [22:18:58] ah, cool [22:25:54] bearND dbrant: i'd swear the aspect ratio on that $50 fire is slightly skewed. when i look at the circle article on device, it looks slightly oblong. i even put a jar on it for reference. when i take a screen cap it looks fine so i'm thinking the display itself is skewed. [22:42:33] jdlrobson: hey, do you know of some change to the MobileWebSectionUsage schema? [22:43:12] it's producing way more events than usual and breaking things [22:44:31] phuedx: ^ [23:00:54] hi madhuvishy [23:01:18] it's configurable [23:01:20] lemme find the line [23:04:06] dbrant: i finally got an OOM. i had don't keep activities on so maybe it's not that surprising. [23:09:37] madhuvishy: https://gerrit.wikimedia.org/r/263538 will need to be swatted. Apparently it wasn't configurable [23:09:40] ^ phuedx [23:11:33] jdlrobson: we are getting the data etc, but the tables will become unquery-able soon. EL itself is fine - master to analytics store replication is super slow though [23:12:59] it was stuck on this schema from yesterday (although not sure if it was this schema causing it or the replication script itself) [23:13:34] but - this is super high throughput, if you want these many events, it would be good to tell us in advance so we can blacklist it from going to mysql and you can read it from hadoop [23:14:26] madhuvishy: i'm a little surprised it spiked so much tbh and haven't had a chance to look closely in detail at what's going on. [23:14:37] jdlrobson: aah [23:14:49] but we made some changes the week before the deployment cut off [23:14:54] and i expect some how that led to more events [23:14:58] it's sending about 130 events per second [23:17:22] or 120 - grafana is a little confusing [23:17:22] do you know how much was it doing before the deployment cut off? [23:17:41] https://grafana.wikimedia.org/dashboard/db/eventlogging [23:17:57] when was the deployment cut off? [23:19:47] spike on 17th [23:20:28] so yeh makes sense. Underestimated the impact on traffic. [23:20:31] jdlrobson: oh wow that's a long time [23:20:45] we're done with it though for the time being [23:20:54] jdlrobson: okay [23:20:55] if there is anything your end you can do to deny events go for it [23:21:02] otherwise my patch above will disable event logging [23:21:10] if we re-enable it we'll have to pick a lower sample rate [23:21:12] ya we can blacklist it from going into mysql if you want [23:21:21] and it'll just be in hadoop [23:21:41] and then you can lower sampling rate and we can revert [23:24:56] jdlrobson, madhuvishy: i've tested the patch locally and, after figuring out what was wrong with my local setup, it lgtm [23:25:02] madhuvishy: as a general rule what's the maximum events per second we should be seeing on that graph? 40? [23:29:19] jdlrobson: finding out on #analytics [23:34:54] jdlrobson: it looks like there's no straight number we can say - it depends a lot on the volume of data being collected, and for how long. [23:38:14] jdlrobson: phuedx - when will the code to lower sampling rate be deployed? should we blacklist for now? [23:38:32] madhuvishy: blacklist first and we'll let you know when it's deployed [23:38:52] (it's equivalent, given our solution) [23:38:59] but ours is less severe :) [23:39:35] phuedx: alright then