15:59:15 #startmeeting metrics team 15:59:15 Meeting started Thu Feb 28 15:59:15 2019 UTC. The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:15 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:59:30 notirl: did you figure out the storm issues? 15:59:34 i'm afraid i can't see/edit the agenda today ): 15:59:41 ok. 15:59:44 irc is working fine though 16:00:21 let's start with topics that don't need a gaba. maybe she'll join later. 16:00:25 ok 16:00:30 OONI response and Sponsor 13 report (karsten) 16:00:43 did you see leonid's response? 16:01:02 i didn't see the latest one, but i'm hoping that it does contain everything we need to do the comparison? 16:01:33 it's from 1 week ago. 16:01:40 did you receive it? 16:01:48 was it the pointer to the github repo? 16:01:54 ooni-pipeline 16:02:19 it did contain a pointer to the github repo, plus lots of details in the email itself. 16:02:31 ah ok, then this is probably everything we need then 16:02:41 that was my impression/hope, yes. 16:02:44 the last one i read only had the github link, but maybe i missed the details 16:02:51 i've been focussing on the first half to get that done 16:02:56 okay. 16:03:16 i need to find a better way of representing the dependencies and data flow because the graph is horrible 16:03:18 this email contains you as recipient, but just in case you're missing details, let me know, and I'll forward this email to you. 16:03:23 ok cool 16:03:35 hi! 16:03:38 hi gaba! 16:03:40 hi gaba! 16:03:49 notirl: feel free to share what you have as graph, and I might have ideas. 16:04:01 ok, i can send it when i am home later 16:04:05 sure. 16:04:10 can you add a note to the agenda? 16:04:18 doing that now. 16:04:21 thanks 16:05:13 done. 16:05:22 moving on to the next topic? 16:05:26 ok 16:05:51 Next retrospective (karsten) 16:06:04 just saw that being edited. 16:06:11 I did manage to add a note before it crashed 16:06:24 I think we are Monday 4th 17:00 UTC according to the pattern so far 16:07:02 ok. 16:07:09 there's a question on the pad to move it to 15:00 utc. 16:07:16 that is fine for me 16:07:17 yes, I'm asking :) 16:07:24 I have a meeting 17:30 UTC on Monday 16:07:25 I can do 15:00 utc. 16:07:30 ok, thanks! 16:07:33 yay consensus 16:07:56 done! 16:08:16 Minimal plan for TorDNSEL rewrite without the fancy parts (karsten) 16:08:25 we carried that over from last week. 16:08:50 context was that we didn't get the grant, and we need to start working on this anyway. 16:09:14 what was the date when the current service gets shut down? 16:09:26 end of march i think 16:09:30 ugh 16:09:49 if we stop producing exit lists, i assume onionoo still serves the latest exit list it has? 16:09:57 welllllll 16:10:19 things will get bad really soon. 16:10:29 we shouldn't let this happen. 16:10:49 so this is a priority 16:10:54 yes. 16:10:56 ok, we should get some tool that produces exit lists running then 16:11:18 so, 16:11:32 silly question, who is running the current service and why it is being shutdown? 16:11:35 i wonder if there is an easy workaround though, can we just run the current code on otf cloud instead of tpa infra? 16:11:43 ^ that 16:12:15 if it's a vm maybe even tpa can give me the image and i just run it somewhere until we're ready to switch over 16:12:29 hmmm :( 16:12:30 ideally we want to run both solutions in paralell to ensure coverage is equal/better 16:13:07 can we go through the list of other priorities for march? 16:13:28 my top priority once sponsor13 is done is exit lists 16:13:29 maybe we can get this done properly in march. 16:13:53 there were interruptions in feb, but i'm hoping sponsor13 is done by the end of next week 16:14:03 so 6 points available for exit lists in march 16:14:06 yes, it seems to me that nothing else (other than s13 report) is top priority right now 16:14:29 in the backlog we have #24422 16:14:37 #25776 16:14:43 ok. let's keep that plan to run existing code somewhere as backup plan. 16:14:58 information architecture is trickling along 16:15:05 onionperf work is ongoing 16:15:33 regarding the exit list scanner, how can I help? 16:15:50 right now this plan relies a lot on your availability to do it. 16:16:02 because you want to reuse existing code, I think. 16:16:26 karsten: we need to extend exit list formats to include ipv6 addresses, source software+version, source identifier 16:16:36 you could start with these spec changes 16:16:53 and once agreed on them you can implement in metrics-lib 16:17:01 maybe also onionoo 16:17:12 okay. 16:17:22 adding that to my list. 16:17:38 notirl: is this something that can be done in march? 16:18:02 yeah, the spec stuff can happen now 16:18:07 we know what we need from it 16:18:16 in brussels we said that needs to be replaced by march 25th 16:18:19 ok 16:18:35 why the 25th? 16:19:02 maybe because it gave us time to overrun? 16:19:10 I guess 16:19:19 the deadline we have on this is march 25th 16:19:23 ok. 16:19:35 do we have a ticket for this? 16:19:42 for the spec change? 16:19:46 i have a post-it to make the tickets 16:19:49 ok 16:19:50 and the metrics-lib/Onionoo implementation? 16:19:50 which i did not do yet 16:19:51 sounds good 16:20:02 ah, for the replacement? 16:20:17 i have post-its for all the tickets that i need to make 16:20:22 ok. 16:20:26 including spec changes, onionoo, metrics-lib, etc. 16:20:33 great! 16:20:46 we need some bulk ticket creation method because it's a little painful 16:20:50 which is why i didn't do it yet 16:20:59 but i'll just do it tonight 16:21:17 ok. all noted in the pad. 16:21:23 thanks 16:21:33 moving on? 16:21:39 yeah 16:21:50 Monthly report (karsten) 16:21:58 we didn't have a monthly report for january. 16:22:18 mostly because things were a bit hectic with brussels. 16:22:29 I wonder, did anybody notice? 16:22:49 it is good to think about what i've been doing even if no one is going to read the report 16:23:03 so, my plan was to write a report for january and february. 16:23:11 ok, sounds good 16:23:30 shall we aim to post this after our retrospective? 16:23:36 sounds good 16:23:43 okay, can you send me a quick update what you have been doing in january _and_ february then? 16:23:54 yeah, will do tomorrow 16:24:07 I can compile a draft by next monday. 16:24:07 we can include some of the stuff that was discussed at the hackweek in the report? 16:24:44 yeah 16:25:01 yes, I can go through my meeting notes and see what makes sense to include. 16:25:44 thanks! at least the roadmap that came out of it 16:26:01 you mean a link to the kanban thing? 16:26:39 yes, there is a public read only link in the wiki page for the metrics page 16:26:47 https://storm.torproject.org/shared/TsgBadLSfM8uh_aftPjxmR_rm0a8E-4UQ2HEa_l0DHv 16:27:06 ok. 16:27:11 makes sense. 16:27:22 alright, moving on? 16:27:29 yes 16:27:39 User-perceived performance, next steps (karsten) 16:27:48 I stopped working on this, even though I could do more. 16:28:01 current status is that I identified some issues in failed and very slow runs. 16:28:22 I posted my findings to the tor-scaling@ mailing list. 16:28:43 by issues you mean stuff that metrics and/or other teams should be working on, right? 16:28:48 there was some discussion with dgoulet and asn on the mailing list, but nothing happened from there on. 16:29:11 I was talking with Isabela aobut next steps and that is why I called for a meeting for tor scaling to put togther priorities/issues we need to work on. 16:29:21 right. that meeting makes a lot of sense. 16:29:21 When is the meeting? 16:29:24 mikeperry also sent a mail with his thought son that 16:29:29 things are stalling at the moment. 16:29:30 there is a doodle to choose a time 16:29:33 let me look for it 16:29:45 https://doodle.com/poll/sym3a3e9cfd5ce99 16:29:52 ok cool 16:30:02 I could also do more by looking closer at slow runs. 16:30:10 but I didn't want to spend even more time on this. 16:30:33 in particular after finding lower hanging fruit that are still on the tree (or however that metaphor goes). 16:30:38 I would love to have a summary of all the discussions somewhere for this meeting and I will try to do that soon. 16:30:50 hehe 16:31:13 does this summary that I'm trying to give here make sense to you? 16:31:21 yes 16:31:29 yes for me :) irl? 16:31:38 yes 16:31:42 cool! :) 16:31:46 so, regarding next steps, 16:31:54 this ties into our earlier discussion of high priorities. 16:32:04 we'll have to do other things first. 16:32:21 yeah 16:32:27 I think I shouldn't work on this stuff in march. 16:32:36 this stuff == user-perceived performance. 16:32:49 but more work remains. 16:33:07 participating in that meeting works just fine. but I cannot do much more on the analysis side. 16:33:20 hope this makes sense. 16:33:35 yes, it makes sense 16:33:40 i'd want to see and participate in the discussion but also i don't think i would have time to write code for it 16:33:49 let's wait until we agree with others on next steps for this 16:33:52 ok 16:33:55 ok. 16:34:40 next? 16:34:43 yeah 16:34:47 yep. just added a note. 16:34:51 Tor is presenting projects to outreachy - https://www.outreachy.org/ 16:35:02 this looks like a different kind of gsoc. 16:35:04 please vote in the doodle <-- last thing about that 16:35:04 maybe next year 16:35:08 yes 16:35:13 outreachy is focus on diversity 16:35:15 in tech 16:35:20 and is also about mentoring 16:35:21 sounds very cool! 16:35:31 but I'm with notirl here. 16:35:36 we just turned down gsoc. 16:35:37 yes, it is ok 16:35:42 i think also, do we not have to pay for outreachy ourselves/ 16:35:44 i was just bringning it in case there was anything 16:35:45 it's twice per year, right? 16:36:01 yes, sara is looking for sponsors for it 16:36:05 ok 16:36:15 let's reconsider in 6 months? 16:36:18 next time. pili is coordinating it and will mention next time 16:36:20 yes, sounds good 16:36:23 ok 16:36:24 cool! 16:36:48 Roadmap: 16:36:57 The last topic is the roadmap and I think you all already answered the question what you are working on 16:37:07 #29507 16:37:13 and sponsor 13 tech report 16:37:15 shall we look at this in the retrospective? i can't see the roadmap at the moment 16:37:30 ok 16:37:32 very quickly: 16:37:39 I'm not working on #29507 anymore. 16:37:44 just to check if there was anything else you are working on 16:37:45 that's something for the icebox. 16:37:45 ohh 16:37:49 ok 16:37:59 it's the user-perceived performance thing. 16:37:59 sorry 16:38:00 yes 16:38:03 no worries. :) 16:38:04 you just said that before 16:38:08 from memory, i think none of my things have changed status 16:38:20 notirl is still working on the reports. 16:38:33 ok 16:38:44 * gaba needs to run to take kids to school... 16:38:45 right now I'm working on #29425, and I wanted to work on #29166. 16:38:57 and now I'm adding the spec thing as even higher priority. 16:39:02 we can talk about this on monday! 16:39:09 ok, thanks! 16:39:16 o/ 16:39:18 great! thank you! ttyl, bye! :) 16:39:21 bye! 16:39:38 #endmeeting