14:02:29 <karsten> #startmeeting metrics team
14:02:29 <MeetBot> Meeting started Thu Oct 27 14:02:29 2016 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:29 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:02:38 <karsten> https://pad.riseup.net/p/3M7VyrTVgjlF <- agenda pad
14:05:25 <karsten> iwakeh: alright, we have a few agenda items there. shall we start?
14:05:30 <iwakeh> yes.
14:05:35 <karsten> * Berlin November meeting (karsten)
14:05:47 <karsten> so, there will be ooni folks in berlin.
14:05:53 <iwakeh> fine.
14:05:55 <karsten> and they're organizing a hackday.
14:06:08 <karsten> I'm considering to join them. are you around on the 11th and/or the 12th?
14:06:20 <karsten> (we're meeting on the 10th)
14:06:43 <iwakeh> Unfortunately, I planned an offline-time 11-13 long ago.
14:06:45 <karsten> somewhat related, we should think about tasks for new volunteers.
14:06:51 <karsten> ah, I see.
14:06:59 <iwakeh> Sure, I can prepare tasks.
14:07:04 <karsten> cool!
14:07:29 <iwakeh> did people already
14:07:51 <iwakeh> state their intention for volunteering?
14:08:13 <karsten> I guess it's mostly an adhoc thing.
14:08:25 <iwakeh> ;-)
14:08:39 <iwakeh> Anyway, there should
14:09:20 <iwakeh> be some tasks available when planning the further CollecTor tasks before/while switching to Onionoo.
14:09:31 <karsten> sounds good!
14:10:00 <karsten> okay, moving on.
14:10:04 <karsten> * Metrics graphs performance improvements (karsten)
14:10:21 <karsten> good news: the graphs showing user numbers are ~3 seconds faster by now.
14:10:38 <karsten> the reason is a switch from read.csv() to load().
14:10:48 <iwakeh> good!
14:10:50 <karsten> https://metrics.torproject.org/userstats-relay-country.html
14:10:52 <karsten> for example.
14:11:17 <karsten> I'm not planning to do more there for now.
14:11:33 <karsten> okay, moving on:
14:11:35 <karsten> * CollecTor 1.1.0 release (#18910, #20380) (iwakeh)
14:11:39 <karsten> the big (!) topic.
14:11:44 <iwakeh> right!
14:11:53 <karsten> did you see my latest comment on #18910?
14:12:00 <iwakeh> no
14:12:05 * iwakeh looking it up
14:12:08 <karsten> take a quick look..
14:12:32 * karsten runs into crazy bugs here.
14:12:44 <karsten> (maybe I should test this stuff more often.)
14:13:23 <iwakeh> yes!
14:13:39 <iwakeh> And write tests for these :-)
14:13:51 <iwakeh> work around?
14:13:55 <karsten> probably.
14:14:07 <karsten> and fix soon after for the next metrics-lib version.
14:14:19 <iwakeh> definitly.
14:14:24 <karsten> I can start a new test run later today.
14:14:33 <karsten> and I'll look into #20380.
14:14:47 <karsten> if all goes well we can release tomorrow. :)
14:14:47 <iwakeh> does the logging work now, for you?
14:14:58 <karsten> I didn't try it yet. was busy with #18910.
14:16:39 <karsten> okay, anything else on this topic that is easier to discuss here than on the tickets?
14:16:45 <karsten> (will be reach 100 comments?....)
14:16:51 <karsten> we*
14:17:00 <iwakeh> surely :-)
14:17:17 <iwakeh> just five more to go
14:17:39 <iwakeh> should have been many separate tickets.
14:18:00 <karsten> :)
14:18:08 <karsten> okay, moving on?
14:18:15 <iwakeh> fine.
14:18:20 <karsten> * Web design tasks Atlas/Metrics (karsten)
14:18:27 <iwakeh> new designer.
14:18:33 <karsten> we'll have a web designer for atlas and metrics in nov and dec. yay!
14:18:54 <iwakeh> what is with the other candidates?
14:18:57 <karsten> some things we cannot start yet, because we'll first want to do an analysis.
14:19:15 <karsten> other candidates for what?
14:19:20 <iwakeh> irl?
14:19:31 <iwakeh> or will both work?
14:19:44 <karsten> I'd say we talk to him before changing anything.
14:20:04 <karsten> pretty sure he wouldn't mind leaving a few tickets to somebody else.
14:20:14 <iwakeh> fine.
14:20:25 <iwakeh> better more than too few people.
14:20:42 <karsten> right. and maybe he can help with reviews.
14:20:51 <iwakeh> good idea.
14:21:04 <karsten> and it would be good to have him for long-term maintenance.
14:21:10 <karsten> so, not planning to push him out in any way.
14:21:32 <iwakeh> just want to make sure I didn't miss anything there ;-)
14:21:33 <karsten> the other code base is the metrics website.
14:21:44 <karsten> sure. please ask if anything seems unclear. :)
14:21:55 <iwakeh> all fine.
14:22:28 <karsten> though we'll want to talk about the metrics website in berlin first.
14:22:49 <karsten> so, maybe there's not much to work on for a web designer before berlin anyway.
14:23:01 <karsten> but we should start thinking about tasks where we'd need one.
14:23:12 <iwakeh> ok
14:23:28 <karsten> okay, that's all on that topic from me. moving on?
14:23:38 <iwakeh> fine.
14:23:42 <karsten> * planning: consider keeping the Onionoo release simple (karsten)
14:23:55 <iwakeh> just the first release
14:24:07 <karsten> we might consider leaving improvements for later, to have more time on all the other tasks.
14:24:28 <iwakeh> agreed.
14:24:34 <karsten> we might even do those improvements in the Sponsor X time frame.
14:24:37 <karsten> but optionally.
14:24:44 <karsten> ok.
14:24:58 <iwakeh> could be more than one release, in case
14:25:02 <iwakeh> we have time.
14:25:03 <karsten> but, I sneaked that item in after you wrote down yours.
14:25:05 <karsten> yes!
14:25:09 <karsten> * planning: consider starting with little-t-tor stats analysis in Dec (karsten, iwakeh)
14:25:37 <iwakeh> This fits in
14:25:40 <karsten> we can always start earlier with something.
14:25:48 <karsten> but should we reflect that in the roadmap?
14:25:54 <iwakeh> right.
14:25:57 <karsten> by changing the end date?
14:26:21 <iwakeh> or mentioning that it'll be
14:26:27 <iwakeh> on time for 030.
14:26:59 <karsten> so, the issue is that we're not only removing code.
14:27:15 <karsten> if we want to obfuscate statistics as they're reported, we'll have to write new code.
14:27:22 <karsten> for: 2017-04: Obfuscate data that gets reported by Tor relays and bridges to the directory authorities by implementing at least one suggestion from the earlier analysis document. (Sponsor X 4.3. Tor daemon)
14:27:32 <iwakeh> Than we should start beginning Dec.
14:27:38 <iwakeh> then
14:27:39 * karsten finds the 030 roadmap..
14:27:54 <iwakeh> net-team page
14:27:59 <irl> wrt atlas, i'm very happy to have someone else do the design and UX work
14:28:02 <karsten> Jan 24th, 2017 is feature freeze.
14:28:12 <karsten> irl: oh, hi! glad to hear.
14:28:27 <iwakeh> what about reviewing?
14:28:44 <irl> if there is a new person coming along, we can discuss splitting up the tickets
14:29:25 <karsten> irl: sounds great! that person has not started yet. I'll tell him to join us here, or set up a separate meeting for that.
14:29:32 <karsten> and let you know, too. :)
14:29:42 <irl> i need to go through and triage again, as i lost a whole load of time due to personal issues
14:29:48 <irl> things are starting to be back on track now though
14:30:30 <irl> i have ooniprobe packaging and atlas triage and maybe small merges planned this weekend
14:31:02 <irl> (including merging my branch with some of the UX fixes)
14:31:05 <karsten> great! let us know if we can help with anything, including review of small/easy patches.
14:31:12 <irl> will do
14:31:48 <karsten> oh, and just in case you happen to fix all the bugs this weekend, we'll have the other person work more on the metrics website. ;)
14:32:04 <iwakeh> hehe
14:32:05 <karsten> okay, back to 030?
14:32:18 <iwakeh> yes, we start ux
14:32:29 <iwakeh> usability analysis
14:32:35 <iwakeh> in november
14:33:09 <karsten> I mean,
14:33:10 <iwakeh> thus, beginning on dec can be for 4.3
14:33:18 <karsten> we don't *have* to get this stuff into 030.
14:33:26 <iwakeh> ah
14:33:28 <karsten> we can also aim for 031.
14:33:37 <karsten> feature freeze May 1st, 2017.
14:33:57 <karsten> we could even do a mix: write the analysis and remove some code for 030, do the bigger changes for 031.
14:34:21 <iwakeh> we could see how many changes and how tricky they are.
14:34:26 <karsten> yep.
14:34:27 <iwakeh> right
14:34:51 <iwakeh> next?
14:34:52 <karsten> okay, so we start the analysis in november and leave the roadmap as it is?
14:34:57 <iwakeh> yes.
14:35:04 <karsten> ok.
14:35:14 * karsten is behind with writing agenda items. what did I miss...
14:35:19 <karsten> err, action items.
14:35:55 <karsten> I wonder who should make a start with that analysis.
14:36:07 <iwakeh> which?
14:36:15 <karsten> I know the code and stats better, but maybe it's better to start without that knowledge.
14:36:20 <karsten> the tor stats analysis.
14:36:38 <iwakeh> ok, I can look at it soon.
14:36:40 <karsten> (it will be november when we talk next week.)
14:36:51 <iwakeh> yes, I know :-)
14:37:12 <karsten> okay!
14:38:29 <karsten> we ran out of agenda items, but we could talk about action items from last week.
14:38:49 <iwakeh> there bugs to be removed...
14:38:58 <iwakeh> fine.
14:41:03 <karsten> regarding the log levels and bad input items,
14:41:12 <karsten> I think we can do those in november.
14:41:26 <iwakeh> have tickets for discussion?
14:41:30 <karsten> we have good things for the october report, no need to generate more in a rush.
14:41:35 <karsten> yes, sounds good.
14:41:35 <iwakeh> true.
14:41:55 <iwakeh> and there has been quite some testing/fixing work cycles.
14:42:17 <karsten> yep
14:42:54 <iwakeh> #20430
14:43:08 <karsten> ok.
14:44:43 <iwakeh> #20421
14:45:49 <iwakeh> what about #19831
14:46:24 <iwakeh> relates to #8799
14:46:42 <iwakeh> and #20323
14:46:55 <karsten> what about those?
14:47:45 <iwakeh> ok, just leave as is, but the compression could be turned on with 1.1.0 anyway?
14:48:18 <karsten> you mean move #19831 to 1.1.0?
14:48:31 <karsten> if so, why not.
14:48:36 <karsten> it just changes a default.
14:48:39 <iwakeh> yes.
14:48:48 <iwakeh> one ticket less to track.
14:48:51 <karsten> want to move to that milestone?
14:48:52 <karsten> :)
14:49:01 <karsten> okay, I hear somebody else wants to give an update.
14:49:13 <iwakeh> hi
14:49:38 <tjr> yea
14:49:40 <tjr> tiny one
14:49:43 <karsten> hi!
14:49:57 <tjr> I've been updating https://consensus-health.torproject.org/ a bit lately
14:50:10 <tjr> The latest thing I've added has been graphs: https://consensus-health.torproject.org/graphs.html
14:50:11 <karsten> great! thanks for taking care of that.
14:50:35 <tjr> If people have suggestions, file a ticket and assign it to me (there's no component for it though)
14:50:51 <karsten> should we fix that?
14:50:55 <karsten> by creating a component?
14:50:58 <tjr> Sure, if you can
14:51:08 <karsten> yep. name?
14:51:16 <tjr> 'Consensus Health'
14:51:22 <karsten> Metrics/Consensus Health ?
14:51:24 <tjr> (the code is named depictor but no one knows that)
14:51:26 <tjr> Sure
14:51:28 <karsten> right.
14:51:33 <karsten> okay, will create that after the meeting.
14:51:42 <karsten> nice graphs!
14:51:58 <iwakeh> they don't show?
14:52:01 <tjr> Secondly, I am still running two collector instances (one on the internet one privately), although they've pretty much unmaintained. I'm excited about this syncing though thing though and promise that at some point I will update these and get the public one syncing if you want it
14:52:08 <tjr> You can't be on High Security slider in TB
14:52:10 <tjr> They're SVG based
14:52:16 <iwakeh> ok ;-)
14:52:33 <tjr> I copied the same graph library from Atlas thinking it'd be better to try and standardize on a single library across tor
14:52:55 <karsten> glad to hear you're still running collector! it would be great if you could run the 1.1.0 release when it's out.
14:53:06 <iwakeh> yes1
14:53:15 <karsten> it should be a lot less painful to run than previous versions. ;)
14:53:23 <iwakeh> true :-)
14:53:27 <tjr> Yea, I am still beflore 1.0.0 - I have your 1.0.2 announcement in my inbox as a todo heh
14:53:30 <tjr> That sounds good :)
14:53:37 <iwakeh> just wait for 1.1.0
14:53:40 <karsten> yep.
14:54:01 <karsten> want to try that and let us know here or via the team list how that works?
14:54:24 <tjr> Running 1.1?
14:54:28 <iwakeh> yes.
14:54:28 <karsten> yes.
14:54:43 <tjr> If it's ready I'll try and switch to it in the next week?
14:54:51 <karsten> should be ready next week.
14:54:54 <tjr> Okay, sure
14:55:01 <iwakeh> fine!
14:55:04 <karsten> thanks. :)
14:55:17 <karsten> okay, plenty of action items.
14:55:26 <karsten> anything else for today?
14:55:28 <tjr> I have a small alerting thing I set up that I use: so it seems my collector instance is still kinda-running. https://github.com/tomrittervg/checker/blob/master/samplejobs/MetricsChecker.py
14:55:40 <tjr> It's pulling the consensus every day at least not sure about all the descriptors
14:56:04 <karsten> the code has changed a lot in the past ~6 months.
14:56:21 <iwakeh> also the configuration.
14:56:44 <karsten> oh, right. best to look at the (new) INSTALL.md and go from there.
14:56:50 <karsten> not just from past experience.
14:57:09 <tjr> sure, I'll just set it up from scratch probably :)
14:57:15 <karsten> yep!
14:57:16 <iwakeh> well, you can
14:57:22 <iwakeh> keep the data though.
14:57:30 <tjr> sure
14:57:51 <iwakeh> And there'll be more freedom where
14:58:08 <iwakeh> to place the data and web-pages and logs.
14:59:20 <karsten> okay, running out of meeting time here.
14:59:32 <iwakeh> yep.
14:59:40 <karsten> ah, next week we'll have run out of summer (time).
14:59:51 <iwakeh> one hour earlier?
15:00:13 <karsten> probably.
15:00:25 <iwakeh> should work.
15:00:29 <karsten> okay, will announce that.
15:00:43 <iwakeh> back to coding.
15:00:48 <karsten> !
15:00:57 <iwakeh> Bye, bye.
15:01:02 <karsten> bye!
15:01:05 <karsten> #endmeeting