16:01:19 <GeKo> #startmeeting network-health 10/25/2021
16:01:19 <MeetBot> Meeting started Mon Oct 25 16:01:19 2021 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:19 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:01:28 <GeKo> hello everyone
16:01:50 <hiro> hello
16:01:53 <GeKo> time for a new [3~ network health meeting
16:01:54 <hiro> o/
16:02:03 <irl[m]> hello
16:02:09 <georgia[m]> hello!
16:02:12 <GeKo> pad is over at http://kfahv6wfkbezjyg4r6mlhpmieydbebr5vkok5r34ya464gqz6c44bnyd.onion/p/tor-nethealthteam-2021.1-keep
16:02:17 <cohosh> hey!
16:02:33 <GeKo> please add things where needed (updates/discussion items etc.)
16:02:54 <georgia[m]> GeKo: newbie question — is there a link that works outside of Tor Browser?
16:03:03 <GeKo> yes
16:03:31 <GeKo> https://pad.riseup.net/p/tor-nethealthteam-2021.1-keep
16:04:08 <gaba> hi!
16:04:21 <GeKo> o/
16:05:07 <georgia[m]> GeKo: thank you!
16:05:17 <GeKo> ggus: you around by chance as well?
16:05:36 <ggus> here
16:05:41 <GeKo> nice
16:06:24 <GeKo> i guess we can start while other folks are still adding things to the pad
16:06:47 <GeKo> maybe with you georgia[m] and irl about the website-redesgin?
16:06:51 <GeKo> *redesign
16:07:01 <GeKo> if so, please go ahead
16:07:24 <georgia[m]> GeKo: Yes — big thing there is that irl you should be getting a contract from us ASAP. it's ready, but I think just needs to get sent to you!
16:08:08 <irl> work has started on this looking at reimplementing the ggplot code using matplotlib, using the current statistics module to give the input data
16:08:13 <GeKo> ncie
16:08:26 <georgia[m]> We have been working on the synthesis and starting on design work. we shared a document out over email, and need to add some more to that and get into the design work
16:08:55 <GeKo> irl: do you have a ticket to track that work or a set of tickets?
16:09:00 <hiro> \o/
16:09:01 <irl> it would be good to see some of the design because right now we end up with just a copy of the original graphs but made with python instead of R
16:09:03 <GeKo> or some other place where one can follow along?
16:09:46 <irl> I think once we see the survey data and design information we will be able to put together specific tickets and we can track those in GitLab
16:09:56 <georgia[m]> irl: yes! we can get into a shared design doc this week
16:10:07 <irl> Awesome :)
16:11:11 <hiro> nice
16:11:16 <gaba> it would be great to have user stories around the use of the data too
16:11:18 <GeKo> sounds good (to both of what you said)
16:11:30 <gaba> not sure if georgia[m] that is something you are doing
16:11:38 <georgia[m]> yes — we'll get that all organized into the document we started
16:11:52 <georgia[m]> we have content in a few disparate places at the moment and are unifying
16:12:30 <SusanKennedy[m]> Preliminary findings from survey: https://docs.google.com/document/d/1G4O-aROaIoSOmZF8KQoIT2oo7AVn9iyZGNcB3wrJsaY/edit?usp=sharing
16:12:49 <GeKo> are you sending  further lnks around or is all supposed to be in the google doc at the end?
16:12:54 <georgia[m]> SusanKennedy[m]: We will keep adding to this document so that there's a hub of information and documents
16:13:20 <georgia[m]> and can update on email as we add to it
16:13:32 <GeKo> wfm
16:13:46 <GeKo> (works for me)
16:13:48 <hiro> sounds good
16:14:24 <irl> Definitely not promising anything but it looks like visualising some data on a map would be a good idea…
16:14:42 <GeKo> georgia[m]: do you need some input/feedback for the stuff you currently have in the google doc?
16:14:54 <GeKo> or is that just a fyi?
16:15:01 <georgia[m]> irl: I could see a map as a good idea too :) but we just need to be careful what we show on the map, as people sometimes think maps say more!
16:15:20 <georgia[m]> GeKo: I think it's ok right now, let us add more and then questions/feedback will be good
16:15:25 <georgia[m]> it's very high level at the moment!
16:15:36 <hiro> something that I will be looking forward would be understanding what is the best way to present the data - I mean understanding this from the users survey
16:16:22 <hiro> something that I would intuitively expect (and I could be wrong) is that we have a set of people that will be looking for network data, a set of people that would be looking at geopolitical events and a set of people that will be looking at ways to link those
16:16:54 <georgia[m]> yes — exactly.
16:16:55 <gaba> and also for us to keep an eye on the network health and work on censorship responose
16:18:22 <GeKo> georgia[m]: is all of the data you have evaluated in the spreadsheet
16:18:33 <GeKo> or is there stuff getting added to that one as well?
16:19:23 <georgia[m]> that is all the survey results that we are using. We have interview synthesis that's in a different document and we'll add it to the main google doc
16:19:43 <GeKo> okay, thanks
16:19:58 <georgia[m]> only thing we may do in the spreadsheet is some of the additional analysis that we mention in the google doc
16:20:11 <georgia[m]> but not new data
16:21:09 <GeKo> alright, do we have anything to add to that sync? anything that needs further discussion for the work in the week ahead?
16:21:24 <GeKo> or can we move on to the next topic?
16:21:49 <irl> I don’t have any thing else
16:22:14 <georgia[m]> all good here!
16:22:21 <GeKo> great
16:22:42 <GeKo> ggus: so, the eol work, how are we doing here?
16:23:00 <GeKo> i keep polling new relay ops to contact and write mails
16:23:16 <GeKo> and then act on replies where needed
16:23:41 <GeKo> i think the main question i have is one we think we should stop and just kick the remaining relays out
16:24:03 <ggus> i still need to reply to one or two operators
16:24:28 <GeKo> i am not really happy to wait until 0.4.7 is stable and picked up by the dir-auths
16:24:59 <GeKo> so relays on old versions vanish "automagically"
16:25:07 <ggus> how many EOL relays we have now?
16:25:16 <GeKo> because that will likely take nother couple of months
16:26:18 <GeKo> let's see
16:27:31 <GeKo> we started with 912 (12.24% of the network by weigth) like a month ago
16:28:06 <GeKo> and are now at 537 (6.43% of the network by weight)
16:29:02 <ggus> do you want to give people one more month to upgrade and then in december we remove them?
16:29:22 <GeKo> back then when we kicked out < 0.2.9 we rejected more relays
16:29:58 <GeKo> with almost twice the weight
16:30:13 <GeKo> so it would not be too bad doing it now
16:30:16 <GeKo> but, yes
16:30:32 <GeKo> i think giving another 4 weeks or so sounds good to me
16:30:41 <GeKo> and then kicking the remaining relays out end of november
16:31:19 <ggus> or if you want to do sooner, it's okay too.
16:31:46 <GeKo> i'll think a bit about it
16:31:58 <GeKo> the sooner the less work is on my plate :)
16:32:05 <GeKo> and we can close a ticket
16:32:09 <GeKo> win-win ;)
16:32:17 <ggus> hah ok!
16:32:20 <ggus> wfm
16:32:33 <GeKo> i thought about writing an update to the tor-relays@ thread
16:32:40 <GeKo> giving the final date for the kick out
16:33:01 <GeKo> maybe a bunch of operators get their act together that way
16:33:13 <GeKo> i can do that this week
16:34:02 <ggus> november 22, before thanks giving
16:34:13 <GeKo> wfm
16:34:17 <ggus> so people can use their holiday to upgrade their relay
16:34:47 <GeKo> wait you mean kicking them out on that day?
16:35:20 <ggus> yes, but it was just a suggestion
16:36:10 <GeKo> okay
16:36:36 <GeKo> the thing is kicking them out on puts their fps on our reject list
16:36:45 <GeKo> and makes it *harder* to upgrade
16:37:04 <GeKo> because they need to change fps, too, or need to contact us getting the fps off our list
16:37:33 <GeKo> so, ideally, i guess that holiday for upgrading is happening before we kick them :)
16:37:42 <GeKo> but i'll think about it
16:37:54 <GeKo> maybe *after* thanksgiving would be a right time then
16:38:00 <ggus> yes
16:38:07 <GeKo> anyway, that's all i have for that topic
16:38:22 <GeKo> do we have anything else to discuss today?
16:38:40 <GeKo> apart from a new meeting time potentially?
16:39:01 <hiro> is it because of the  winter time mess?
16:39:04 <hiro> xD
16:39:14 <hiro> the new meeting time...
16:39:23 <GeKo> yes
16:39:41 <ggus> GeKo: if we have some minutes, we could discuss the new 'run a bridge' campaign?
16:39:42 <GeKo> eu is moving the upcoming weekend
16:39:56 <GeKo> ggus: please
16:40:11 <GeKo> i commented on the ticket this morning
16:40:20 <GeKo> but if there is stuff to discuss, go ahead
16:41:25 <ggus> would be possible to have a grafana dashboard to monitor the bridge campaign?
16:41:51 <GeKo> i suppose so, hiro?
16:41:54 <hiro> what do we need on the dashboard?
16:42:56 <ggus> when the new bridge joined, if it's running, and if it has obfs4
16:43:15 <hiro> on a table?
16:43:18 <ggus> yes
16:43:33 <hiro> ok
16:43:41 <hiro> yes can do
16:43:59 <hiro> do we want also - I don't know - number of bridges measured per day?
16:44:32 <hiro> I think I just need everything that we want and I can start putting together a dashboard this week
16:44:46 <ggus> okay! i will create a ticket for you!
16:44:49 <hiro> as a first iteration
16:44:53 <irl> Watch out for what running means
16:45:04 <hiro> yeah we have a discussion about that
16:45:44 <hiro> https://gitlab.torproject.org/tpo/network-health/team/-/issues/131
16:46:34 <irl> Yeah that is probably orports being disabled
16:46:50 <GeKo> not alays
16:46:56 <GeKo> *always
16:47:24 <GeKo> ggus: anything else you need from us?
16:47:28 <GeKo> or want to chat about?
16:47:51 <ggus> i will reply to your comment. but i'm good
16:47:57 <GeKo> thanks
16:48:08 <GeKo> okay, meeting time
16:48:18 <GeKo> do we keep 1600utc?
16:48:27 <GeKo> or do we move to 1700utc and if so, by when?
16:48:38 <GeKo> curently the network-team is meeting 1700utc
16:48:49 <hiro> I like to keep time in UTC and adjust but some people like the opposite
16:48:52 <GeKo> and they'll have a discussion about that in 10 minutes
16:49:27 <hiro> I guess we should follow what the previous and next meeting do
16:49:42 <hiro> but I hpe they agree on the same
16:49:47 <GeKo> :)
16:49:57 <GeKo> okay, i'll sort that out with the browser folks
16:50:07 <GeKo> and the network-team people
16:50:28 <GeKo> and hopefully i'll announce something this week on the network-health list
16:50:46 <GeKo> meanwhile let's keep it 1600utc for next monday?
16:51:56 <GeKo> sounds like, "yes" :)
16:51:58 <GeKo> okay then
16:52:07 <GeKo> any last minute items?
16:53:00 <GeKo> nothing. have a nice week then everyone
16:53:03 <GeKo> #endmeeting