14:00:06 <karsten> #startmeeting metrics team
14:00:06 <MeetBot> Meeting started Thu Jul 14 14:00:06 2016 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:06 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:13 <karsten> hi, who's here for the meeting?
14:00:25 <iwakeh> me :-)
14:00:26 <anathema_db> me
14:00:35 <karsten> hi iwakeh and anathema_db!
14:00:46 <karsten> please add topics to the pad.
14:00:53 <karsten> https://pad.riseup.net/p/zUNzEIFRq5S4
14:00:57 <anathema_db> hi karsten
14:02:31 <anathema_db> I confess: I didn't read last meeting's minute
14:02:43 <iwakeh> no problem
14:02:44 <karsten> no worries. :)
14:03:06 <anathema_db> but I'd like to know more about the python style guide and if there are any news on the Berlin meeting
14:03:19 <iwakeh> well,
14:03:28 <iwakeh> I'll add it to the pad.
14:03:34 <anathema_db> k thanks
14:04:33 <anathema_db> I've just added one small thing
14:04:38 <anathema_db> it's more of a question
14:04:48 <karsten> sure.
14:04:59 <karsten> okay, 5 minutes into the meeting, let's start with agenda item 1.
14:05:03 <karsten> * MOSS award start date (karsten)
14:05:12 <karsten> we're free to pick a date of our choice, and we picked july 1.
14:05:20 <iwakeh> ok.
14:05:29 <karsten> so, the award will run from july 1, 2016 to june 30, 2017.
14:05:36 <karsten> which is good, I think.
14:05:45 <iwakeh> yes.
14:05:53 <karsten> nothing more about that. paperwork still in the making.
14:06:05 <anathema_db> cool
14:06:16 <karsten> moving on to...
14:06:17 <karsten> * Berlin meetup (karsten)
14:06:18 <anathema_db> there will be any kind of new team?
14:06:24 <karsten> hmm?
14:06:36 <anathema_db> new organisation, new people working on the MOSS award items?
14:06:46 <karsten> ah, right.
14:06:53 <anathema_db> or we'll spread the workload among us ?
14:07:38 <karsten> that's still a bit unclear until the paperwork is signed, except for iwakeh's work which is already part of the award.
14:07:48 <anathema_db> ok, great
14:07:58 <karsten> but we'll have to think about what we do with the remaining funding.
14:08:24 <karsten> or rather, make decisions. we already thought a bit about this a few weeks ago.
14:08:26 <anathema_db> yeah I guess we need to address how to use them
14:09:11 <karsten> so, related to this, I put up the deliverables a week or so ago.
14:09:14 <karsten> let me find them.
14:09:26 <anathema_db> m
14:09:27 <karsten> https://trac.torproject.org/projects/tor/wiki/org/sponsors/SponsorX
14:09:28 <anathema_db> *k
14:10:01 <anathema_db> ok
14:10:06 <anathema_db> read it, thanks
14:10:35 <karsten> I hope we'll add more content to that wiki page soon. iwakeh, let's think about that this week, ok?
14:10:50 <iwakeh> right, I started tht a little.
14:11:18 <iwakeh> should we link back to the sponsorX page?
14:11:29 <karsten> from other pages? sure.
14:11:41 <iwakeh> ok, fine.
14:11:45 <karsten> or from that page to others? sure, that, too.
14:12:01 <karsten> alright. Berlin!
14:12:06 <karsten> https://trac.torproject.org/projects/tor/wiki/org/meetings/2016BerlinMeetup
14:12:11 <karsten> we have a weekend.
14:12:16 <iwakeh> well ...
14:12:23 <anathema_db> oh so we already have a date?
14:12:23 <anathema_db> good
14:12:52 <karsten> we're not going to spend the full three days on metrics stuff.
14:12:54 <karsten> ;)
14:12:58 <anathema_db> eheeh
14:13:03 <iwakeh> maybe we should find more detailed time frames?
14:13:19 <iwakeh> reserve the space
14:13:35 <iwakeh> for metrics only topics? or some space?
14:13:56 <karsten> we could start by picking 1 of the 3 days for metrics stuff.
14:14:09 <iwakeh> but not on the current meeting.
14:14:16 <iwakeh> maybe doodle again?
14:14:28 <karsten> yes, sure.
14:14:39 <karsten> I'll start a doodle for that.
14:14:45 <anathema_db> sorted
14:14:46 <iwakeh> great!
14:15:12 <karsten> so, if you're planning to attend at least one day, would you mind putting your name on the wiki page?
14:15:18 <anathema_db> can someone put the address of the Onion Space?
14:15:21 <karsten> right now it looks like we'll be three people.
14:15:27 <karsten> unclear!
14:15:33 <karsten> I don't know how public it's supposed to be.
14:15:40 <anathema_db> ah ok, got it
14:15:43 <karsten> it's in Berlin Wedding.
14:16:06 <iwakeh> four people ;-)
14:16:09 <karsten> but I'll have to ask them if they want the address to be public. otherwise we'll share by email.
14:16:10 <anathema_db> well, I'm in the middle of changing company so I've to double check if I can take Friday off
14:16:14 <karsten> yay, it's growing!
14:16:17 <anathema_db> will update later in the month
14:16:34 <karsten> okay, I'll make that doodle a yes/no/maybe.
14:16:51 <karsten> so, anything else about berlin?
14:17:05 <anathema_db> no I'm ok
14:17:11 <karsten> * documentation for projects (iwakeh)
14:17:12 <iwakeh> Development documentation:
14:17:20 <iwakeh> proposed new section on team page with links to project pages:
14:17:30 <iwakeh> https://trac.torproject.org/projects/tor/wiki/org/teams/MetricsTeam#Projects
14:17:38 <iwakeh> just place holders
14:17:54 <iwakeh> to get an idea; more in CollecTor
14:18:24 <karsten> hmm, would it be too much to include what specifically the project will be about?
14:18:37 <iwakeh> Well, that's the ToDo
14:18:49 <karsten> oh okay.
14:19:28 <iwakeh> Is the list ok?
14:19:46 <iwakeh> I mean: "This list is not exhaustive, but rather lists some long-term ..."
14:19:50 <anathema_db> should we add Atlas?
14:19:58 <iwakeh> Sure.
14:20:01 <karsten> probably, yes.
14:20:28 <karsten> but, are we talking about the same definition of project here?
14:20:38 <anathema_db> mm
14:20:39 <karsten> defined start and end date, can fail, etc.
14:20:42 <iwakeh> That's what I try to find out.
14:20:44 <karsten> rather than code base.
14:20:53 <iwakeh> Ah,
14:21:08 <iwakeh> you intended a project plan ? Gannt chart?
14:21:20 <karsten> well, I can imagine two things:
14:21:29 <karsten> 1. a list of actively maintained code bases,
14:21:42 <iwakeh> agreed.
14:21:44 <karsten> 2. a list of active projects to improve some code base in some way.
14:22:14 <iwakeh> I think, for 2. we'd have to do too many wiki page updates.
14:22:15 <karsten> where a project could even affect two code bases (onionoo change and corresponding atlas extension),
14:22:26 <karsten> and there could be 0 or more than 1 projects for the same code base.
14:22:40 <karsten> okay, so 1.
14:22:48 <anathema_db> I think that could be confusing so maybe it's better to stay with 1.
14:23:12 <iwakeh> thanks, that's what I indended to find out
14:23:18 <iwakeh> before writing more.
14:23:28 <karsten> is there a better word for project then?
14:23:29 <iwakeh> Therefor the TODOs.
14:23:34 <iwakeh> codebase
14:23:46 <iwakeh> is fine, but technical.
14:23:56 <anathema_db> I think project is fine
14:23:57 <karsten> in the past I used product or codebase, I think.
14:24:11 <anathema_db> product sounds like corporate
14:24:14 <anathema_db> :)
14:24:16 <iwakeh> product or areas of work
14:24:17 <karsten> well, many people say project, but that's confusing.
14:24:27 <karsten> with other projects that start, end, fail, err.
14:24:39 <anathema_db> code project?
14:25:03 <iwakeh> Systems
14:25:04 <karsten> agree with the "code" part. ;)
14:25:17 <anathema_db> code project systems.
14:25:18 <anathema_db> lol
14:25:19 <karsten> hmm, https://www.torproject.org/getinvolved/volunteer.html.en#Projects says projects, too.
14:25:35 <karsten> okay, just a thought, no need to find a winner now.
14:25:47 <anathema_db> which makes sense
14:25:48 <karsten> darn, didn't we want to finish in 30..
14:25:55 <iwakeh> I'll leave it as project for now?
14:25:58 <karsten> ok.
14:26:20 <iwakeh> the versioning could wait
14:26:28 <karsten> okay, moving it down.
14:26:32 <karsten> * python style guide next steps (iwakeh)
14:26:48 <karsten> (guess we can go until :45 like last time)
14:27:02 <iwakeh> well, I intend to propose PEP8 to metrics-ml
14:27:12 <anathema_db> I agree with PEP8
14:27:25 <iwakeh> if noone yells, that's it and later addition are wellcome, of course.
14:27:36 <karsten> sorted. :)
14:27:41 <iwakeh> ok.
14:28:03 <karsten> it sounds like the right decision.
14:28:11 <karsten> without having written much python, that is.
14:28:12 <iwakeh> yes, it reads fine
14:28:21 <iwakeh> the guide PEP8
14:28:23 <iwakeh> and is
14:28:31 <iwakeh> close to what we use in java.
14:28:38 <karsten> cool.
14:28:41 <iwakeh> as far as possiblee.
14:29:01 <karsten> okay, moving forward?
14:29:06 <anathema_db> k
14:29:08 <iwakeh> yep.
14:29:10 <karsten> * Onionoo addinational infos (anathema_db)
14:29:24 <anathema_db> yeah so you may already know what I'm talking about
14:29:32 <iwakeh> protocol?
14:29:48 * karsten hasn't read that thread in detail yet.
14:29:49 <anathema_db> I'd like to know if there are any stats on the service/code
14:29:54 <karsten> ah.
14:30:07 <karsten> you mean specifications of the machine?
14:30:10 <anathema_db> no more about system usage, benchmarks
14:30:21 <iwakeh> users/hour?
14:30:30 <anathema_db> like ram consumption, cpu workload under avg/max load
14:30:44 <karsten> https://munin.torproject.org/torproject.org/omeiense.torproject.org/cpu.html
14:30:50 <qiv> all data science python i have seen was pep8, great choice :-)
14:30:51 <anathema_db> have you in mind apachebenchmark output?
14:30:56 <karsten> user tor-guest, password anything
14:31:07 <iwakeh> @qiv, thanks.
14:31:34 <anathema_db> karsten: is that specific to Onionoo's server?
14:31:45 <karsten> anathema_db: omeiense is onionoo only.
14:31:46 <anathema_db> I'm more interested in how much resource Onionoo process takes
14:31:51 <anathema_db> cool
14:31:58 <anathema_db> and how about benchmarks?
14:32:16 <anathema_db> like time to respond to a request, amount of data returned, etc etc ?
14:32:19 <qiv> iwakeh: there is one exception though, tensorflow, they use google ... but anyways, i love pep8
14:32:37 <karsten> anathema_db: there's the stuff that gets logged.
14:32:53 <karsten> anathema_db: I might have shared that in the past, but let me find a recent output.
14:33:37 <anathema_db> you shared some data that was extracted from the code itself
14:33:40 <karsten> anathema_db: http://paste.debian.net/781283/
14:33:41 <anathema_db> with a specific class
14:33:54 <karsten> yes, that's custom code.
14:34:04 <iwakeh> logging, actually
14:34:13 <anathema_db> yeah I''m more interested in benchmarks
14:34:25 <karsten> that's all we have, I think.
14:34:26 <iwakeh> that's one form of benchmark.
14:34:33 <anathema_db> something like:
14:34:39 <anathema_db> (waiting for link...)
14:35:17 <iwakeh> I used Jmeter long ago for Onionoo, but not data anymore ...
14:35:36 <anathema_db> https://paste.debian.net/781284/
14:35:55 <anathema_db> yeah I may get into jmeter as well, just need to study it
14:35:58 <anathema_db> I'm more familiar with ab
14:36:06 <anathema_db> but that's the kind of data I'm interested on
14:36:29 <anathema_db> so that's just 1 request
14:36:34 <iwakeh> I think Jmeter is quite good for that purpose.
14:36:43 <anathema_db> it is yes
14:37:10 <iwakeh> But that also measures network and other outside parameter.
14:37:21 <anathema_db> so I'd like to know what is the consumption of RAM and CPU when under load
14:37:27 <anathema_db> even more for multiple requests
14:37:55 <anathema_db> yep, more a real case scenario (simulating a client)
14:37:56 <karsten> the log statements I pasted were just from the java part that answers web requests.
14:38:13 <karsten> there's also a varnish cache now that prevents many, many requests from ever reaching the java part.
14:38:19 <karsten> and there's of course network overhead.
14:38:31 <karsten> I'm open to adding more measurements. should we do that on trac?
14:38:36 <anathema_db> ah ok good for varnish
14:38:38 <iwakeh> yes.
14:38:45 <anathema_db> can you explain also the architecture?
14:38:54 <karsten> of onionoo?
14:39:01 <iwakeh> which detail?
14:39:22 <iwakeh> maybe, per mail?
14:39:24 <anathema_db> architecture
14:39:25 <karsten> ah, you mean, some overview that includes the word "varnish"?
14:39:30 <anathema_db> yeah sure, mail is fine
14:39:48 <karsten> sounds good.
14:39:58 <anathema_db> like the java service, then nginx (example), varnish cache
14:39:59 <anathema_db> etc etc
14:40:09 <anathema_db> the db
14:40:17 <anathema_db> and so on
14:40:18 <karsten> yes! want to ask on the list, and iwakeh or I respond?
14:40:32 <iwakeh> or both.
14:40:35 <anathema_db> sounds good
14:40:43 <karsten> great!
14:41:02 <karsten> anything else about onionoo?
14:41:17 <karsten> by the way, feel free to move the discussion back to the mailing list.
14:41:25 <anathema_db> perfect, thanks
14:41:28 <anathema_db> I'm good for now
14:41:31 <anathema_db> we can move forward
14:41:35 <karsten> ok.
14:41:38 <karsten> * release process, i.e. use versions or tags or (iwakeh)
14:42:08 <iwakeh> for bundling issues in a release versions would be neat.
14:42:11 <karsten> tags are easiest, but we can work with other trac features.
14:42:23 <iwakeh> tags also cause mails. and
14:42:36 <iwakeh> you could just create a batch of new versions in on ticket?
14:42:59 <iwakeh> I think we should sort early
14:43:00 <karsten> so, hmm.
14:43:15 <iwakeh> e.g. descriptor 2.0.0 and 1.4.0 and ...
14:43:22 <karsten> I think tor uses versions to say which version contains a bug.
14:43:28 <karsten> not which version should contain the fix.
14:43:37 <iwakeh> ah, ok.
14:43:37 <karsten> it uses milestones for that.
14:43:47 <iwakeh> well, that's fine with me
14:43:54 <karsten> and I can create versions and/or milestones as needed.
14:44:06 <karsten> I should create a ticket and list them, then create them, then close the ticket. easy.
14:44:09 <iwakeh> milestones has a nice overview
14:44:19 <iwakeh> ok. fine.
14:44:33 <karsten> do you have a list?
14:44:45 <karsten> or just those two for descriptor?
14:44:46 <iwakeh> in my head :-)
14:44:49 <karsten> heh.
14:45:06 <iwakeh> and for the others the very first
14:45:23 <iwakeh> releases and also the next ones for
14:45:32 <iwakeh> fixes and mini features.
14:45:51 <iwakeh> we could make a list
14:46:01 <iwakeh> in the new ticket before you create them.
14:46:04 <karsten> hmm, on second thought, I don't think we need to document creation of new milestones and versions. just components.
14:46:12 <karsten> but having a ticket wouldn't hurt.
14:46:14 <iwakeh> even better
14:46:27 <karsten> to create this list.
14:46:37 <iwakeh> right.
14:46:48 <karsten> want to create a ticket when you have a good first list?
14:46:55 <iwakeh> ok ;-)
14:47:15 <karsten> or, send me the list, and I create the ticket?
14:47:22 <iwakeh> also fine.
14:47:32 <iwakeh> I'll mail the list.
14:47:33 * karsten thinks of other people seeing the ticket and going ahead with creating things.
14:47:37 <karsten> great!
14:47:40 <karsten> oookaaaay
14:47:49 <karsten> anything else for today?
14:47:54 <anathema_db> I'm good
14:47:54 <iwakeh> all done.
14:48:03 <karsten> wonderful. thanks for a great meeting!
14:48:10 <anathema_db> thanks to you guys
14:48:14 <karsten> talk to you in a week.
14:48:17 <iwakeh> thanks. bye bye.
14:48:18 <karsten> #endmeeting