15:00:13 <karsten> #startmeeting metrics team
15:00:13 <MeetBot> Meeting started Thu Mar  2 15:00:13 2017 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:13 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:21 <karsten> hello :)
15:00:42 <RaBe> hey :)
15:01:07 <iwakeh> 13 seconds late ;-)
15:01:20 <iwakeh> Meeting started Thu Mar  2 15:00:13 2017 UTC
15:01:44 <karsten> yeah, sorry for that :D
15:02:10 <karsten> okay, shall we start? anything missing in the agenda?
15:02:20 <iwakeh> sure.
15:02:45 <karsten> okay.
15:02:47 <karsten> - Atlas usability project remaining fixes and blog post (irl|RaBe)
15:02:49 <irl> hi
15:03:04 <irl> - 16 tickets closed this week
15:03:06 <karsten> amazing to see all those closed trac tickets!
15:03:15 <RaBe> :D
15:03:17 <irl> i've had to keep updating the agenda for this number
15:03:24 <irl> because patches just kept coming
15:03:28 <irl> and they were all good
15:03:32 <iwakeh> great!
15:03:36 <karsten> awesome! :)
15:03:41 <irl> we're getting through these quite quickly now
15:04:11 <irl> there's less than 30-ish left i think
15:04:26 <irl> 22 to be exact
15:04:33 <karsten> maybe the blog post will summon more contributors.
15:04:44 <irl> indeed
15:05:01 <irl> we have at least one, maybe more than one cypherpunks that has been very helpful
15:05:15 <karsten> good to hear.
15:05:26 <irl> on the topic of new contributors, i've prepared a wiki page
15:05:34 <irl> - New wiki page with information for prospective contributors https://trac.torproject.org/projects/tor/wiki/doc/Atlas
15:05:55 <irl> there is a hacking section here that should help new contributors get started, improvements to this are welcome
15:06:10 <iwakeh> cool!
15:06:28 <karsten> cool! mind if I look at the blog post and the wiki page after the meeting?
15:06:38 <iwakeh> (TODO: Which mailing list should be used for Atlas?)
15:06:39 <irl> that seems ok
15:06:53 <karsten> (which would be effectively tomorrow, because of another meeting after this one.)
15:07:05 <irl> karsten: no problem
15:07:09 <karsten> fine question about the mailing list.
15:07:20 <iwakeh> On the bottom of the page.
15:07:32 <irl> i was thinking tor-relays@ for users and metrics-team@ for dev
15:07:45 <irl> unless you think it's worth a dedicated list
15:07:46 <karsten> that could work.
15:07:51 <irl> (which i don't)
15:08:10 <karsten> I'd say let's use existing lists and see how that goes.
15:08:28 <iwakeh> metrics-team is good as things are related to Onionoo often.
15:08:31 <karsten> maybe mention that both lists are public.
15:08:46 <irl> ok, i'll do that
15:09:38 <karsten> sounds great!
15:09:41 <irl> (:
15:09:57 <irl> unless RaBe has anything futher on Atlas, I think that concludes Atlas for now
15:10:01 <irl> oh, except one thing
15:10:32 <irl> this is part of a Sponsor X deliverable, which mentions the scope for bugs, which should probably be used to prioritise bugs
15:10:46 <irl> can someone at some point point me at the missing context i need to work out what they are?
15:11:05 <iwakeh> high-priority are the focus for X.
15:11:12 <karsten> oh, I think we're doing just fine with respect to sponsor X.
15:11:23 <karsten> high priority by 2016-12-31.
15:11:29 <iwakeh> right.
15:11:41 <irl> ok cool. i just stumbled across it in the wiki and wondered if there was anything that was important
15:11:48 <irl> that sounds fine
15:11:50 <karsten> well, we mostly put that in, so that we don't suddenly have to solve all tickets (which you almost did ;)).
15:12:02 <irl> heh (:
15:12:03 <iwakeh> we should have known :-)
15:12:07 <karsten> true!
15:12:18 <iwakeh> overachieving is good :-)
15:12:37 <davywybiral> Hi all. Does Tor Browser cache view-source: pages?
15:12:51 <karsten> yes, we did a lot more than we had to. but that's okay.
15:13:01 <karsten> users actually use this stuff.
15:13:15 <karsten> davywybiral: in a meeting right now. want to ask in #tor-project or come back in 45 minutes?
15:13:29 <karsten> alright, shall we move on?
15:13:40 <RaBe> sure :)
15:13:53 <karsten> - CollecTor, Onionoo, and metrics-lib websites (karsten|iwakeh)
15:14:03 <karsten> there are at least 3 topics here:
15:14:07 <karsten> - move sub sites to Metrics
15:14:07 <karsten> - split Metrics into data-processing service and website
15:14:08 <karsten> - rebranding (or let's wait postpone this until it's on Trac)
15:14:21 <karsten> oh, and another one:
15:14:34 <karsten> - metrics-lib contents
15:14:44 <iwakeh> web-site?
15:14:45 <karsten> all somewhat related.
15:14:52 <karsten> - metrics-lib website contents
15:15:20 <karsten> so, iwakeh and I just had a discussion of the metrics-lib website.
15:15:21 <iwakeh> well, the last is going forward.
15:15:28 <karsten> and we have a plan.
15:15:42 <karsten> our action items are:
15:15:42 <karsten> iwakeh:  https://trac.torproject.org/projects/tor/ticket/21614
15:15:43 <karsten> karsten: write a draft page of what we discussed above
15:16:00 <karsten> which I'll share on the ticket next week.
15:16:20 <iwakeh> 3 needs some discussion via mail still?
15:16:42 <karsten> probably.
15:17:10 <iwakeh> Is 2 in scope for q1&q2 2017?
15:17:28 <karsten> not necessarily.
15:17:44 <karsten> it would be good to discuss whether it's a good idea though.
15:17:54 <karsten> but the actual work can happen after sponsor X.
15:18:17 <iwakeh> was there a ticket?
15:18:24 <karsten> for 2?
15:18:34 <iwakeh> yes?
15:19:01 <iwakeh> ah, right.
15:19:04 <karsten> #19754
15:19:53 <karsten> I wonder how to move forward here. hmm.
15:20:00 <iwakeh> I think I can respond to the comment there.
15:20:29 <iwakeh> Currently, web and backend are intertwined.
15:20:50 <karsten> irl: would you want to comment on #21551 and whether that would work for you?
15:20:56 <karsten> iwakeh: true.
15:21:12 * irl takes a look
15:21:21 <karsten> irl: basically, you brought up concerns last week, and I hope we addressed those.
15:22:03 <karsten> okay, seems like we can continue on trac here, and hopefully discuss this more with a trac ticket for 3 next week.
15:22:18 <iwakeh> ok.
15:22:29 <irl> karsten: all the concerns are there, so that's ok
15:22:41 <karsten> irl: okay!
15:22:54 <karsten> moving on?
15:23:00 <iwakeh> yep.
15:23:14 <karsten> hiro: around?
15:23:21 <karsten> uhhhhm
15:23:23 <karsten> skipped one.
15:23:27 <karsten> - Tech reports on statistics collected by the Tor daemon (iwakeh)
15:23:34 <karsten> (sorry!)
15:23:44 <iwakeh> Aiming at next week new draft
15:23:58 <karsten> yeah, sorry for keeping you busy with all the other stuff..
15:24:11 <iwakeh> well,
15:24:17 <iwakeh> 'aiming'
15:24:24 <hiro> yes i am here
15:24:25 <iwakeh> but, I might have
15:24:31 <hiro> sorry :)
15:24:38 <iwakeh> a list of questions for you by tomorrow.
15:24:44 <karsten> sounds good!
15:25:58 <karsten> and just in case there's a question that you need an answer to more quickly, don't wait for the list to be finished. ;)
15:26:12 <iwakeh> good to know :-)
15:26:21 <karsten> okay, moving on?
15:26:35 <iwakeh> yes.
15:26:40 <karsten> - Deployed OnionPerf instances (hiro)
15:26:44 <hiro> yep
15:26:45 <karsten> hiro: there.
15:27:01 <hiro> so I was making an issue when parsing the time differences
15:27:28 <hiro> I thought I was getting the difference in microseconds instead I was just getting the microseconds part of the difference
15:27:35 <karsten> I noticed the new graphs look more similar to torperf than the old ones.
15:27:42 <hiro> so I fixed that and the data in the board made more sense yes
15:27:50 <karsten> great!
15:28:11 <hiro> I am going to add to the documentation the tpo machine
15:28:18 <hiro> I will send that to the ticket too
15:28:25 <hiro> so we will have 3 instances in the end
15:28:33 <karsten> cool!
15:28:56 <hiro> onionperf-one which is the US-DC instance.. then onionperf-eu which is in amsterdam and the onionperf.tpo
15:29:02 <hiro> s/the/then
15:29:27 <iwakeh> where is onionperf.tp.o?
15:29:27 <hiro> that's my update
15:29:35 <karsten> hetzner?
15:29:36 <iwakeh> just curious.
15:30:12 <karsten> so, hmm, we're printing these names out on the metrics website.
15:30:22 <karsten> I wonder if there's a way to make them more intuitive to users.
15:30:40 <karsten> maybe onionperf-$cc for all of them?
15:30:55 <hiro> $cc ?
15:30:59 <karsten> country code.
15:31:05 <hiro> ah ok
15:31:11 <hiro> yes we can do that
15:31:54 <karsten> or op-$cc ?
15:32:06 <karsten> not sure if onionperf-$cc is too long.
15:32:11 <hiro> the question is do you want to do that at the host level, onionperf logs or collector?
15:32:43 <karsten> host level would be easier, I guess.
15:32:49 <iwakeh> op-$cc might be better.
15:32:50 <hiro> ok :)
15:33:02 <karsten> otherwise different collector instances could collect these logs differently.
15:33:15 <hiro> ok
15:33:15 <iwakeh> they could always
15:33:17 <karsten> I think they shouldn't mess with the source name.
15:33:25 <karsten> they could, but not accidentally.
15:33:36 <iwakeh> true.
15:33:49 <karsten> hope that's not too much trouble, hiro?
15:33:53 <iwakeh> but, the source is configuration.
15:34:09 <iwakeh> so, operators will mess with it ;-)
15:34:13 <hiro> it isn't :)
15:34:21 <karsten> yes, but onionperf operators.
15:34:31 <karsten> not collector operators.
15:34:45 <iwakeh> ah, I was referring to collector ops.
15:35:58 <karsten> hmm, not in the branch I'm working on, which I never showed anyone. ;)
15:35:59 <karsten> +## OnionPerfHosts = http://first.torproject.org/,http://second.torproject.org/
15:36:17 <karsten> collector just uses whatever source onionperf puts in.
15:36:23 <karsten> in that branch. anyway. :)
15:36:38 <iwakeh> ok,
15:37:06 <iwakeh> I thought it would be analogue to torperf there.
15:37:25 <karsten> ah, you think it's a feature. hmm.
15:37:31 <karsten> maybe.
15:37:39 <karsten> lots of things to discuss when that branch is ready.
15:37:40 <iwakeh> well, we might want to change names
15:37:54 <iwakeh> and researchers might need it
15:38:03 <iwakeh> for their own perf-files.
15:38:29 <iwakeh> yes :-)
15:38:46 <karsten> yes, let's take that to the ticket once there's something that works. which is not the case yet
15:38:55 <karsten> okay, cool!
15:39:08 <iwakeh> I need to write text anyway :-)
15:39:13 <karsten> I hear that sponsor R folks are waiting for this, too.
15:39:43 <karsten> shall we move on to the last topic?
15:40:14 <iwakeh> sure.
15:40:19 <karsten> - Monthly team report (karsten)
15:40:26 <karsten> I put the current draft on the pad.
15:40:34 <karsten> anything that needs tweaking before that goes out?
15:41:21 <iwakeh> fine from my point of view.
15:41:25 <karsten> cool!
15:41:30 <irl> atlas bits look fine
15:41:34 <karsten> great!
15:41:40 <irl> yep, all good
15:41:47 <iwakeh> (linebreaks and XXX replacements, of course :-)
15:41:52 <karsten> heh, yes.
15:42:25 <karsten> okay, how about I leave it there for another 15 minutes until 16:00 UTC and then send it out?
15:42:41 <karsten> except there's need to edit something, in which case I don't send it yet?
15:43:35 <iwakeh> main block were agreed upon.
15:43:41 <iwakeh> blocks
15:44:09 <karsten> good!
15:44:12 <irl> (:
15:44:24 <karsten> okay then. done on time? :)
15:44:27 <iwakeh> :-)
15:44:37 <karsten> 3 weeks until amsterdam, by the way.
15:44:45 <hiro> yay!
15:44:46 * Samdney comes too late and lurks :)
15:44:56 <iwakeh> yes.
15:45:10 <karsten> Samdney: any pressing questions before we end the meeting?
15:45:23 <Samdney> not for the meeting
15:45:27 <karsten> okay. :)
15:45:43 <karsten> great, thanks for this meeting, everyone! talk to you next week.
15:45:47 <karsten> #endmeeting