15:05:39 <karsten> #startmeeting metrics team 15:05:39 <MeetBot> Meeting started Thu Jan 16 15:05:39 2020 UTC. The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:05:39 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 15:06:03 <gaba> Let's change the name to 2020 or start a new one for this year. 15:06:06 <gaba> (about the pad) 15:06:13 <karsten> oh, it has a name. 15:06:24 <karsten> what's your preference? 15:06:51 <gaba> I would like us to use a pad.riseup like the other teams and archive the old year in nc. :) 15:07:04 <karsten> do you mind creating a pad? 15:07:18 <karsten> or, we could even do that after the meeting. 15:07:22 <gaba> https://pad.riseup.net/p/tor-metricsteam-2020.1-keep 15:07:31 <karsten> okay. 15:08:45 <acute> I just made it 15:08:48 <acute> hi everyone 15:08:49 <karsten> hi! 15:09:15 <gaba> hi! 15:10:19 <karsten> should we start? 15:10:28 <irl> ok 15:10:51 <karsten> MaxMind database license issue (karsten) 15:11:02 <karsten> so, maxmind changed their license. 15:11:23 <karsten> they require signing up for an account and agreeing to some new license. 15:11:50 <karsten> this affects us with onionoo not having recent geolocation and AS data. 15:12:15 <karsten> it also affects us with tor not being shipped with a recent country database that statistics are based on. 15:12:31 <gaba> is there any alternative to use? 15:12:53 <karsten> nick points out that there's also a debian-legal thread about this: https://lists.debian.org/debian-legal/2020/01/threads.html 15:13:22 <karsten> there are alternatives, but we'd have to evaluate them. 15:13:50 <karsten> we'll have to resolve this in the next weeks. 15:13:57 <irl> we /could/ comply with their license 15:14:12 <karsten> for onionoo? 15:14:47 <karsten> in the onionoo case we do not ship the database. we just use it. 15:14:54 <irl> hmm 15:14:55 <karsten> in the tor case we ship a file based on their database. 15:15:10 <anarcat> looks like debian is talking about just dropping the DB and providing tools for users to fetch the db from maxmind 15:15:14 <karsten> ideally, we'd take the same approach for onionoo and tor. 15:15:17 <irl> yeah in the tor case we can publish a list of ip addresses that tor relays should delete from their geoip 15:16:49 <karsten> how would that work, irl? 15:17:24 <irl> when we receive a complaint from maxmind that we need to remove some ip address we add it to some file on our website 15:17:49 <irl> that then instructs tor clients to destroy those entries from the db 15:18:03 <irl> it's a silly solution but that is the solution if we wanted to comply i think 15:18:38 <irl> it looks like debian have settled on not including the data from reading that thread 15:19:54 <karsten> so, hmm. 15:20:07 <karsten> I'm not sure about implementing such hacks in order to comply. 15:20:08 <anarcat> for what it's worth, there are two core maxmind-related packages in debian: https://tracker.debian.org/pkg/geoip-database - the actual database, which would be removed under the proposed plan, and https://tracker.debian.org/pkg/geoipupdate which is a tool to download the database from maxmind, given the proper credentials, which would stay 15:20:22 <anarcat> this could be delegated to users 15:20:32 <anarcat> at least that's the position debian seems to be heading towards 15:20:33 <karsten> users would be relay operators. 15:20:35 <anarcat> yes 15:20:42 <anarcat> and onionoo operators i guess :) 15:20:43 <karsten> it makes sense for them to do that. 15:20:47 <karsten> yes, that would work. 15:20:58 <karsten> but I don't see how relay operators would actually do this. 15:21:02 <karsten> at least not many of them. 15:21:12 <irl> how many relays need to report stats for our stats to be meaningful? 15:21:55 <karsten> hmmm. 20%? wild guess. 15:22:10 <anarcat> yeah it's a huge shift 15:22:17 <karsten> we include them if we have >=10%, I think, but they're not that useful then. 15:22:24 <anarcat> but it's been a long time coming, i'm surprised maxmind let this slide for so long 15:22:26 <karsten> and how would this work for non-debian relays? 15:22:56 <karsten> they even gave a 2 weeks heads-up... 15:23:29 <karsten> anyway, we're not going to solve this today. I just wanted to point this out as something we'll have to look into. 15:23:40 <karsten> maybe the next step could be to list alternatives? 15:23:42 <gaba> do we have a trac ticket for this? 15:23:48 <karsten> not yet. 15:23:55 <karsten> I can start this. 15:24:01 <karsten> any quick ideas for alternatives? 15:24:02 <irl> there was a paper on k anonymity for ip address data 15:24:22 <irl> if a database was implementing those techniques, it wouldn't be subject to this law 15:24:23 <karsten> I did a comparison of geoip databases many years ago. I could start there. but maybe there are more recent alternatives? 15:24:31 <irl> or at least if law respects maths, that would be the case 15:24:54 <irl> https://arxiv.org/pdf/1707.03900.pdf 15:24:57 <anarcat> law and math rarely go well together unfortunately 15:25:14 <anarcat> i don't think there are alternatives to maxmind so far - the debian-legal thread talks about doing a fork, but i am not sure that is realistic 15:25:41 <karsten> I'm also concerned about the work. 15:26:08 <irl> i notice maxmind has me in the wrong country 15:26:17 <anarcat> honestly, this ip geolocation stuff has always been kind of bullshit :p 15:26:20 <irl> and the wrong isp 15:26:36 <anarcat> here i'm in the wrong province, and in canada that means like 600km+ away 15:26:47 <irl> ip addresses as geo-authentication for copyright access is a great idea 15:27:19 <irl> should we all take an action item to think hard about this? 15:27:32 <karsten> yes, please. 15:27:38 <karsten> I'll also create a ticket. 15:27:43 <karsten> _and_ think hard about it. ;) 15:27:44 <irl> ok cool 15:28:26 <karsten> next topic? 15:28:33 <irl> ok 15:28:46 <gaba> ok 15:28:49 <karsten> Needs review tickets 15:28:58 <gaba> there are 3 there. What is the status? 15:29:37 <karsten> one is in need of review for under a day. that's for whenever irl has time. 15:29:46 <karsten> the others are in need of review by me?! 15:30:11 <irl> i don't think the others needs review 15:30:14 <acute> all onionperf automation is still irl I think :) 15:30:18 <irl> we already discussed these enough that i think they are reviewed 15:30:25 <karsten> okay, cool. 15:30:33 <irl> and the onionperf reviews are not in that list, but i need to review them 15:30:34 <gaba> can that ticket change status then? 15:30:44 <irl> yes 15:31:03 <gaba> ok 15:31:48 <karsten> thanks for watching for this stuff. I think we're good with reviews at the moment. 15:31:57 <karsten> irl did a couple reviews earlier this week, which helped a lot. 15:32:08 <karsten> that's where those four releases came from. 15:32:16 <irl> (: 15:32:23 <gaba> :) 15:32:40 <karsten> okay, roadmap? 15:32:45 <gaba> yep 15:33:06 <irl> for me, no change since yesterday 15:33:10 <gaba> I still need to clean all the info that came out of the roadmap meeting yesterday. I will do it for next week. 15:34:24 <gaba> There is a bunch of stuff in the 'on review' column in trello too. 15:34:33 <gaba> Like #32724 but that trac ticket is not on review. 15:35:40 <acute> yes, these are the tickets I meant! I can change the status for these 15:35:50 <irl> these are in gitlab though? 15:35:53 <gaba> thanks acute 15:35:55 <gaba> ahhh, ok 15:35:57 <gaba> that is why 15:35:58 <karsten> #32747 is closed, because it's merged, but I need to check whether it actually worked. 15:36:06 <acute> these are on metrics-cloud, not onionperf 15:36:07 <irl> ooh, they are in cloud 15:36:09 <irl> i see 15:37:45 <karsten> so, the tasks we brainstormed yesterday will also go into this roadmap? 15:37:54 <gaba> Let's add a comment to the link in gitlab to the ticket. 15:37:57 <gaba> karsten: yes 15:38:14 <karsten> should we then go through the remaining items in this roadmap and decide what to do with them? 15:38:17 <gaba> I will add them to the icebox and we can sort them out into backlog 15:38:22 <gaba> yes, mostly the backlog 15:38:35 <acute> this sounds great! 15:39:00 <karsten> I think we'll also have to clean up in the icebox and the done column at some point. 15:39:37 <gaba> karsten: let's do that in the next time we meet for roadmapping. If there is something that will not happen then it should be removed from icebox and ticket can be closed. 15:39:50 <karsten> sounds good. 15:40:16 <karsten> wait, do you mean next week or so, or in 3 months? 15:41:31 <gaba> after we get the tasks coming out of ipv6. Remember we still need that and the obj 3 for the scalability project. 15:41:40 <karsten> okay, cool. 15:42:15 <karsten> so, in progress and on review are good now? 15:43:37 <karsten> are we done? 15:43:40 <gaba> the on review then is on somebody's plate? 15:43:57 <gaba> #32724, to 27... Just want to confirm that this is the case 15:44:22 <irl> yes i will look at those 15:44:42 <gaba> irl: where are you keeping track of those or where you send reviews/comments? 15:44:45 <acute> and we can move them to done as soon as the code is merged 15:44:57 <irl> on the tickets 15:45:06 <gaba> https://trac.torproject.org/projects/tor/ticket/32724 ? 15:45:10 <irl> acute: did you not have trac setting to review privs? 15:45:14 <irl> or do you have those now 15:45:21 <irl> gaba: yes 15:45:33 <acute> I will check 15:46:47 <gaba> I'm ok with the roadmap then :) 15:47:05 <acute> yes, I can change them :) 15:47:09 <irl> woo 15:47:16 <karsten> not sure if this is a roadmap topic, but I'd like us to work harder on resolving defect tickets. 15:47:55 <karsten> some of these are really old. I started fixing/resolving some of them, but some still remain. 15:48:04 <gaba> karsten: do you have a trac query with priorities? 15:48:29 <gaba> We can take a look at that list in the weekly meetings. 15:48:32 <karsten> we didn't use priorities consistently. 15:49:07 <karsten> https://trac.torproject.org/projects/tor/query?status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&type=defect&component=%5EMetrics&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=priority 15:49:46 <karsten> I could try to prioritize them until next week. 15:49:49 <gaba> karsten: we could go through that list and clean it up 15:49:49 <gaba> yes 15:50:12 <karsten> except for Cloud and Bot I guess. 15:51:57 <gaba> ok 15:52:00 <karsten> alright! 15:52:03 <karsten> that's all? 15:52:08 <gaba> yep 15:52:33 <irl> cool 15:52:53 <karsten> great. have a great evening/day, everyone! talk to you next week. o/ 15:52:58 <gaba> o/ 15:53:03 <acute> bye! 15:53:10 <karsten> #endmeeting