15:01:11 <karsten> #startmeeting metrics team
15:01:11 <MeetBot> Meeting started Thu Apr 25 15:01:11 2019 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:11 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:01:16 <karsten> https://storm.torproject.org/shared/5h1Goax5eNusxjXJ_Ty5Wl7hFR1uqCReUiN8xdlBG8T <- agenda pad
15:01:59 <karsten> who's here for the metrics team meeting?
15:02:12 <acute> hi!
15:02:17 <karsten> hi acute!
15:02:32 <irl> is gaba here?
15:03:18 <gaba> yes
15:03:19 <gaba> hi
15:03:22 <karsten> hi!
15:03:53 <karsten> shall we start?
15:03:56 <irl> ok
15:03:58 <karsten> or add more topics to the agenda?
15:04:20 <gaba> i do not have anything else to add
15:04:53 <irl> no more topics from me
15:04:59 <karsten> me neither. we have a lot of discussions going on in parallel on tickets, but I don't think we need to discuss any of those now.
15:05:25 <karsten> acute: any other topics you want to add?
15:05:30 <gaba> ok
15:05:39 <acute> no
15:05:44 <karsten> okay!
15:05:48 <karsten> -  #29399 (Retire host and services for tordnsel and check)
15:06:00 <gaba> I added that one because weasel was asking
15:06:17 <irl> I said we have deprioritised it until it is funded, which is my understanding
15:06:57 <gaba> we do not have funding for any of the stuff that is happening in metrics. we should do our own priorities right now outside of what is funded or not.
15:07:20 <gaba> i understood it was deprioritized or it is going to be done in the next year, right?
15:07:32 <irl> i think weasel would like it done a lot sooner
15:07:48 <karsten> aren't we going to apply for funding for this work very soon?
15:08:21 <irl> i have lost track of that
15:08:31 <gaba> yes but I think we should work on needed stuff without funding. when we apply then we can change roadmaps according to that.
15:08:46 <karsten> still, I think it's not highest priority for us at the moment.
15:08:54 <gaba> the plan is to schedule a conversation between us and al to talk about that funding.
15:08:57 <gaba> ok
15:08:58 <karsten> if we don't work on the bandwidth file archive soon, we'll lose more data.
15:09:04 <gaba> ok
15:09:15 <karsten> if we don't work on the performance stuff soon, the results will be less relevant.
15:09:30 <gaba> It is fine if that is the priority for that specific reason. We should communicate it with weasel that way.
15:09:47 <karsten> the trouble with backups is new.
15:10:01 <karsten> last thing we knew before that was that we still have 1 year until LTS runs out.
15:10:18 <karsten> and even though we didn't want to use that full year, having at least a few months would be good.
15:10:48 <gaba> ok
15:11:08 <karsten> I see that irl has already answered on the ticket.
15:11:27 <karsten> I only skimmed the conversation there, but that answer sounds reasonable to me.
15:11:47 <gaba> ok. sounds good
15:11:50 <karsten> I'll take a closer look later.
15:12:01 <karsten> and then reply on the ticket.
15:12:34 <karsten> anyway.
15:12:45 <karsten> anything else on that topic?
15:12:48 <gaba> yes, the reason we are not working on it is not because there is no funding
15:12:59 <gaba> but yes, we can continue on the ticket
15:13:33 <gaba> sounds good to me, thanks :)
15:13:52 <karsten> thanks for helping with this, and sorry it's such an annoying topic... :(
15:14:30 <karsten> okay. next topic doesn't need discussion?
15:14:40 <irl> mike's experiments are not going to work
15:14:41 <karsten> it's meant as an announcement only?
15:14:44 <gaba> net one that i wanted to bring for discussion but just a reminder :)
15:14:52 <karsten> hah
15:15:07 <irl> relays count for 24 hour periods, but not necessarily starting at midnight
15:15:12 <irl> they all have different 24 hour periods
15:15:24 <irl> he needs to have 72 hour windows and measure the middle 24 hours
15:15:43 <gaba> irl: can you make comments to mike, in #tor-internal or tor-scalability mailing list or meeting? we need to discuss this. mikeperry is working on this specifically.
15:15:49 <gaba> are you able to come to the meeting on friday?
15:16:03 <irl> possibly as long as it's not clashing with driving lesson
15:16:09 <gaba> it would be good if there is some discussion before that meeting
15:16:14 <irl> otherwise i'll post to tor-scaling
15:16:17 <gaba> thanks!!
15:16:41 <karsten> okay, cool!
15:16:48 <karsten> now, roadmap?
15:16:54 <gaba> yes
15:17:08 * irl loads up roadmap
15:17:22 <acute> unfortunately, the roadmap does not load for me
15:17:24 <gaba> any change on any of the cards in the 'in progress'?
15:17:29 <gaba> hey, there is a bug in wekan
15:17:36 <gaba> click on 'all the baords' at the top
15:17:37 <acute> #29376 can be moved to done
15:17:40 <gaba> and then click on the metrics one
15:17:46 <acute> (if this is still in progress on the board)
15:18:02 <acute> #29374, #29366 are under review
15:18:03 <gaba> yes, thanks
15:18:21 <acute> I mean, they can be moved to 'under review'
15:18:55 <gaba> ok
15:19:23 <karsten> I'm working on #29507 in parallel to #30216 and other children of #21378.
15:19:42 <karsten> but mainly #29507 and #30216.
15:19:57 <karsten> irl took over #29315 from me.
15:20:07 <irl> yes, i am working on #29315
15:20:12 <irl> I think the PRD for data portal is "under review". there is no ticket but antonela is the reviewer.
15:20:51 <gaba> ok, sounds good
15:21:06 <karsten> I wonder what to do with Bridgedb stats.
15:21:14 <karsten> I have a bunch of emails in my inbox on that topic.
15:21:30 <acute> sorry, one more from me : #26673 is in progress (PR ready to submit)
15:21:39 <irl> i think it may be necessary to do them twice
15:21:51 <karsten> but I'm careful not to add yet another thing to my in progress items.
15:22:01 <karsten> do what twice?
15:22:09 <karsten> bridgedb stats?
15:22:18 <irl> we currently don't have the resources to do bridgedb stats in a useful way, we might just let network team do whatever and then later retrofit or rewrite it into our pipeline
15:22:37 <karsten> anticensorship team
15:22:49 <irl> which is part of the network team
15:22:52 <irl> (:
15:22:52 <karsten> and yes, I agree that we don't have the resources at the moment.
15:22:54 <karsten> oh?
15:22:55 <karsten> ok. :)
15:22:59 <irl> i thought it was, maybe i'm wrong
15:23:05 <irl> "incubating"
15:23:08 <karsten> heh
15:23:23 <gaba> karsten: bridgedb stat let's include phw in the conversation
15:23:44 <gaba> anti-censorship team is its own team now that includes cohosh and phw
15:24:07 <karsten> yes, happy to discuss this with phw. my point is that we might not be able to have this conversation in the next 2 or 3 weeks.
15:24:14 <irl> right
15:24:15 <karsten> it's a whole new topic to add these stats.
15:24:25 <irl> it's a context switch which uses lots of mental energy too
15:24:26 <karsten> and we need to focus on this to do it right.
15:24:30 <karsten> yes.
15:24:36 * phw pops in
15:24:39 <karsten> hi!
15:24:45 <phw> hi karsten!
15:25:02 * phw reads backlog
15:25:07 <karsten> we're just talking about delaying the bridgedb stats work for a bit, because we're overloaded with other stuff at the moment.
15:25:34 <karsten> same with snowflake stats, though I don't know how far that has progressed so far.
15:26:02 <gaba> phw: is that blocking any work that needs to be done for s19 ?
15:26:27 <gaba> karsten, ir: funny thing is that anything related to snowflake or bridgedb is funded by s19 :)
15:26:37 <karsten> right!
15:26:40 <phw> i understand.  that's fine from my side.  we can still collect these statistics independently, without directly having them go into tor metrics.
15:26:57 <karsten> are there deadlines?
15:27:16 <gaba> sounds good! i think mostly we need to be sure it is not blocking any work on the anti-censorship team
15:28:14 <phw> it's not blocking anything.  there's still merit in having the data, and publishing some insights from it, even if it's not fully integrated in tor metrics yet.
15:28:31 <gaba> karsten: the sponsor ends at the end of may but there will be other sponsor after that. but it seems we do not need this soon.
15:28:33 <phw> (for bridgedb, that is)
15:28:56 <karsten> ideally, we'd have this discussion before collecting and using data, though. hmmm.
15:29:12 <irl> this is what i'm saying, we could just do it twice
15:29:24 <irl> and bill the funder twice
15:29:36 <gaba> ?
15:29:46 <gaba> what do you mean?
15:29:47 <phw> yes, it would certainly be nice to have someone confirm that the data collection approach is safe.
15:30:04 <irl> phw: can you use the research safety board for this?
15:30:17 <karsten> unclear if that's faster.
15:30:52 <karsten> so, end of may is in ~5 weeks.
15:31:00 <karsten> what needs to happen in that time?
15:31:03 <phw> i'm happy to contact the board if you think that's the best way forward.
15:31:20 <gaba> this sponsor is very flexible. we do not have any specific objective other than work on this
15:31:32 <karsten> but we want to show results at the end of may?
15:32:20 <karsten> stating the obvious: this is not much time.
15:32:32 <irl> gaba: by this logic, the sponsor could also fund the tor metrics work to integrate the stats
15:32:35 <karsten> and if it includes time to collect data, it's almost no time.
15:33:01 <gaba> yes irl
15:33:02 <irl> or at least, tor metrics time to do the design
15:33:13 <irl> which the follow-on sponsor would then fund the implementation of
15:33:22 <karsten> that would work, time-wise.
15:33:34 <karsten> s/would/could/
15:34:14 <karsten> let me suggest something: let us continue/hopefully complete our current tasks in the next two weeks.
15:34:29 <karsten> and then we focus more on the bridgedb (and snowflake?!) thing.
15:34:36 <gaba> sounds good to me.
15:34:49 * gaba needs to run out now but it seems we are fine.
15:35:21 <karsten> ok.
15:35:34 <gaba> o/
15:36:13 <phw> thanks karsten and irl!
15:36:23 <phw> do you want me to approach the safety board in the meanwhile?
15:37:04 <irl> it is not a terrible idea to have a second opinion (or fallback first opinion)
15:37:37 <karsten> what about snowflake?
15:37:55 <karsten> do we need to do something there until end of may, too?
15:38:28 <phw> let me see if cohosh is around.
15:39:21 <karsten> irl: in the meantime, should we move #28322 out of the backlog and into the icebox?
15:39:40 <karsten> and #24422?
15:39:55 <cohosh> hi o/
15:39:58 <irl> #28322 yes, not #24422
15:39:59 <karsten> hi cohosh!
15:40:16 <irl> being realistic with time
15:40:27 <karsten> cohosh: we were just talking about whether we need to do something with snowflake stats by end of may.
15:40:37 <karsten> for sponsor 19.
15:40:42 <cohosh> ah yeah
15:40:51 <cohosh> i was thinking about this again this morning
15:41:08 <cohosh> and i think the best plan is to not try to get all of the stats at once and start with some really simple things
15:41:19 <cohosh> like for example time series information about how many snowflake we have
15:41:30 <cohosh> so that we can compare that to client usage metrics collected from the bridge
15:41:45 <karsten> okay, so we should schedule this for may then.
15:42:00 <cohosh> yeah i have replying to the snowflake stats ticket on my todos for this week
15:42:32 <karsten> so, we're trying to juggle many things at the moment.
15:42:53 <karsten> we might not be able to do anything for this in the next two weeks.
15:43:03 <cohosh> ok that makes sense, i just want to make sure i'm not the one blocking you on this
15:43:16 <cohosh> we're ok if it's not done immediately
15:43:17 <karsten> no, that's not the case.
15:43:24 <karsten> but do you need something by end of may?
15:43:27 <cohosh> cool
15:43:37 <cohosh> the number of snowflakes timeseries would be good to have, yeah
15:43:44 <karsten> ok.
15:44:04 <karsten> thanks!
15:44:17 <karsten> our roadmap might be a tiny bit more accurate now.
15:44:21 <cohosh> np!
15:44:32 <karsten> okay, anything else for today's meeting?
15:44:37 * cohosh also expresses appreciation on the input on safely collecting metrics
15:44:57 <karsten> :) great!
15:45:26 <karsten> irl, acute: anything else for today?
15:45:34 <irl> not from me
15:45:35 <acute> not from me
15:45:54 <karsten> perfect, thanks! talk to you next week then!
15:45:59 <karsten> #endmeeting