[00:46:14] YuviPanda: is your proxy doing regular http proxying for metrics.wikmedia.org? [06:04:58] (PS1) Rfaulk: add - MySQL-python and Flask-Login packs. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84486 [06:04:59] (PS1) Rfaulk: add - more docs around virtualenv. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84487 [06:05:23] (CR) Rfaulk: [C: 2 V: 2] add - MySQL-python and Flask-Login packs. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84486 (owner: Rfaulk) [06:05:33] (CR) Rfaulk: [C: 2 V: 2] add - more docs around virtualenv. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84487 (owner: Rfaulk) [08:07:09] (PS1) Rfaulk: fix - return empty list rather than None. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84494 [08:07:10] (PS1) Rfaulk: add - specify full path to brokers. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84495 [08:07:32] (CR) Rfaulk: [C: 2 V: 2] fix - return empty list rather than None. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84494 (owner: Rfaulk) [08:07:46] (CR) Rfaulk: [C: 2 V: 2] add - specify full path to brokers. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84495 (owner: Rfaulk) [10:37:31] (PS1) Rfaulk: fix - url formatting for hashing. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84512 [10:37:32] (PS1) Rfaulk: mod - use existing data path definition. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84513 [10:37:33] (PS1) Rfaulk: mv - api job queue config to module init. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84514 [10:37:52] (CR) Rfaulk: [C: 2 V: 2] fix - url formatting for hashing. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84512 (owner: Rfaulk) [10:38:01] (CR) Rfaulk: [C: 2 V: 2] mod - use existing data path definition. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84513 (owner: Rfaulk) [10:38:11] (CR) Rfaulk: [C: 2 V: 2] mv - api job queue config to module init. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84514 (owner: Rfaulk) [11:02:27] (PS1) Rfaulk: add - docs on adding new users. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84519 [11:02:51] (CR) Rfaulk: [C: 2 V: 2] add - docs on adding new users. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/84519 (owner: Rfaulk) [15:10:30] yoyo [15:11:33] yo qchris [15:11:36] yo milimetric [15:11:39] yo ottomata [15:11:40] drdee Morning! [15:11:50] had some internet connection issues [15:11:54] now using a hotspot [15:12:04] is milimetric around? [15:24:14] hi! [15:24:18] hey drdee [15:24:35] all the hangouts we make say "the party is over" [15:24:49] yeah we need to update all the linsk [15:25:01] https://plus.google.com/hangouts/_/570a26cbc65ccc87b63fdf1511e9350fe3abb788 [15:25:04] is the link of the meeting [16:59:36] today's scrum will be held in https://plus.google.com/hangouts/_/570a26cbc65ccc87b63fdf1511e9350fe3abb788 [16:59:41] finding a spot! [16:59:50] average: ^^ [16:59:58] tnegrin: ^^ [17:01:19] drdee: I'm not allowed to join [17:01:40] rd> [17:01:48] try again [17:02:10] this client changes the focus when I'm authenticating [17:02:58] old or new bat cave? [17:03:47] https://plus.google.com/hangouts/_/570a26cbc65ccc87b63fdf1511e9350fe3abb788 <-- tnegrin [17:03:56] batcaves are dying [17:04:01] we have to heal them [17:07:11] drdee, we don't have a kafka monitoring card yet, right? if not i will create one [17:07:55] yes i forgot to add that one [17:07:59] please add it [17:08:40] qchris: i will check with dan regarding 1128 [17:08:54] my connection is really spotty [17:14:40] drdee: [17:14:41] https://mingle.corp.wikimedia.org/projects/analytics/cards/1152 [17:14:52] i've been working on that so you might want to move it into this sprint [18:17:04] hey ottomata, dr0ptp4kt [18:17:05] err [18:17:06] drdee: [18:17:06] i mean [18:17:15] yooy [18:17:15] yo [18:17:25] the proxy sets x-forwarded-proto to http or https [18:17:30] so you can use that to redirect http to https [18:22:31] hmmm ok cool [18:23:43] :) [18:24:26] yuvi: [18:24:32] 'X-Forwarded-Proto' like that? [18:25:15] ottomata:yeah [18:25:25] proxy_set_header X-Forwarded-Proto $scheme; [18:28:53] coool [18:28:54] got it [18:28:56] its working better [18:28:58] milimetric: [18:29:09] i think that wikimetrics does some internal redirects or something [18:29:16] and might have a manual 'http' in the redirect urls [18:29:23] i can look [18:29:34] do you have an example? [18:29:39] https://gist.github.com/ottomata/6598610 [18:29:44] 10.4.0.214 - - [17/Sep/2013:18:28:33 +0000] [metrics.wmflabs.org/sid#7f7c1f36d658][rid#7f7c1830f0a0/initial] (4) RewriteCond: input='http' pattern='!https' => matched [18:29:49] that was from an https request [18:30:03] but some internal redirect or request still went http [18:33:29] I don't think so ottomata [18:33:43] it looks to me like it's saying it came in as http [18:33:48] and got rewritten as https [18:34:01] that's before wikimetrics isn't it? [18:34:18] i'll paste the full log of a request [18:35:34] milimetric: https://gist.github.com/ottomata/6598610 [18:35:45] so you can see the first bit of the req [18:35:46] 10.4.0.214 - - [17/Sep/2013:18:35:17 +0000] [metrics.wmflabs.org/sid#7f7c1f36d658][rid#7f7c1830b0a0/initial] (4) RewriteCond: input='https' pattern='!https' => not-matched [18:35:49] means that it came in as https [18:35:58] that's just for GET / [18:36:28] which line is that [18:36:42] 4 [18:36:42] ? [18:36:48] scroll to line 117 [18:37:14] when it requests /login [18:41:55] ottomata: it doesn't look to me like wikimetrics is opinionated about the protocol [18:42:06] because its redirects seem to be relative not absolute [18:42:47] the problem is between these two lines: [18:42:48] (3) applying pattern '^(.*)$' to uri '/login' [18:42:52] (4) RewriteCond: input='http' pattern='!https' => matched [18:43:09] hm yeah but the input is from the X-Forwarded-Proto header [18:43:14] the uri is /login, but something *thinks* it's http [18:43:29] that Yuvi's thing sets up, so it should only be http if the request was made to his proxy as http [18:44:21] what is this &next= stuff? [18:44:24] ?next=http%3A%2F%2Fmetrics.w... [18:44:41] that's telling it where to go after login [18:44:57] I grab the request path and stick it on as a parameter [18:45:11] ohhhHHh [18:45:13] lemme check out that code [18:45:15] yeah, that makes sense then [18:45:25] because the request to wikimetrics looks like http [18:45:31] since it is a proxy request [18:45:42] i guess that's ok, since it will redirect again just like it is doing [18:45:43] hmmm [18:45:56] yea, I'm just looking at request.url [18:46:17] you could get around it be checking the X-Forwarded-Proto too, and using that as the request.url protocol [18:46:29] buuut, maybe that's annoying [18:46:30] and not worth it [18:46:32] I can try to get a relative url too no? [18:46:37] since this extra redirect [18:46:41] it's not actually redirecting though is it? [18:46:45] it's just rewriting [18:47:16] hmmmm i tis redirecting since my rule tells the user's browser to redirect to https [18:47:24] anytime it gets a non-http request [18:47:34] so that tells the proxy to re-do the request at https [18:48:04] the proxy is rewriting, so i think yeah, to the user's browser [18:48:14] it won't really see any redirects (beyond the initial request) [18:48:16] i *think* [18:48:29] anyway, meh, its fine i guess [18:49:25] hmm, works now? [18:49:51] I *could* theoretically add the feature to make the proxy itself do the redirect, but I prefer to not do that to make the proxy as fast as possible [18:50:03] yeah i think the rewrite is good [18:50:27] also, check this out: [18:50:27] wikimetrics.instance-proxy.wmflabs.org [18:50:34] redirects to metrics.wmflab.org now [18:51:18] ok, thanks ottomata. I added a TODO to check the protocol and redirect more intelligently [18:51:52] it also looks like if you're logged out you now get a bunch of errors related to https, so I have to fix that too [18:53:10] yup ! cool [18:53:24] woohoo, all sweet then :) [18:53:43] jaja, danke! [20:49:52] hey drdee [20:49:56] yo [20:50:09] one quick question: [20:50:37] can you pull up the analytics calendar in the My Calendars or Other Calendars list? [20:50:46] yes [20:51:03] interestingly, I see the calendar now *after I create an event* [20:51:11] but I still cannot pull it up from that list [20:51:17] I'll go and talk to chip [20:51:33] k, sorry for the hassle [20:51:46] np - EZ is on it too right? [20:52:13] yes [20:52:17] k cool [21:52:10] ottomata: yes! [22:14:53] qchris_away: check https://mingle.corp.wikimedia.org/projects/analytics