13:59:14 #startmeeting Community Team meeting - April 04 2022 13:59:14 Meeting started Mon Apr 4 13:59:14 2022 UTC. The chair is ggus. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:59:14 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:59:18 hello folks 13:59:20 hello! o/ 13:59:27 community team meeting is starting now! o/ 13:59:30 hello! o/ 13:59:34 o/ 13:59:39 here is our meeting pad: https://pad.riseup.net/p/tor-community-team-2022-keep 13:59:53 please add your updates and feel free to add items on the discussion 14:00:30 o/ 14:00:38 * gman999 lurking and working 14:00:40 Hi! 14:08:03 it seems it will be a short meeting today. 14:08:17 we don't have any topics to discuss 14:08:29 at least on the discussion agenda 14:09:39 do people would like to share or discuss something? 14:10:07 i wanted to point out something 14:10:18 when gathering traffic data for the narrative report 14:10:59 i found an interesting spike in ukraine 14:11:01 https://metrics.torproject.org/userstats-relay-country.html?start=2022-02-01&end=2022-03-31&country=ua&events=on 14:11:19 right around russia's invasion 14:12:02 then the number of relay users went down almost back to pre-war levels 14:12:07 i was wondering what the explanation could be 14:12:44 🤷 14:13:20 I have been thinking it was people being misplaced... having to leave and not having computers 14:13:27 many towns were destroyed 14:13:46 I mean I don't know but that could be an explanation 14:13:53 for sure.. of course 14:14:08 i meant, what could've encouraged the spike, was there a campaign? 14:14:51 maybe the russian soldier were circumventing ukrainian censorship? 14:15:16 unlikely, but i dont have a better explanation 14:15:16 they are mostly not allowed to have mobile phones with them 14:15:27 i read somewhere that russian army had problems to connect to the phone network and get internet 14:15:40 there are massive comms problems for the ru army 14:15:51 it's an interesting event that we don't have one answer 14:15:52 maybe it could also be people from russia using vpns in ucraine 🤷 14:15:54 so there is cell phone usage despite official policy 14:16:24 some days ago people were discussing that topic on #tor-relays channel 14:16:48 i posted two useful URLs in #cake 14:17:22 well, 150000 is more than overall number of russian soldiers in the country 14:17:30 it worth remember that on the week of the invasion there was a massive cyberattack in UA 14:17:50 so maybe that motivated people to download tor? 14:18:30 perhaps 14:18:51 i would want to look closer with PTs, etc. 14:19:06 last year there was another spike in UA that we couldn't find a good answer - https://metrics.torproject.org/userstats-relay-country.html?start=2021-02-01&end=2022-03-31&country=ua&events=off 14:19:33 would ooni ppl know more? 14:19:55 but we thought that was a geoip issue? 14:20:05 gman999: https://metrics.torproject.org/userstats-bridge-combined.html?start=2022-02-01&end=2022-03-31&country=ua 14:20:07 yes, hiro. 14:20:24 ggus: ah interesting.. 14:20:46 so then we updated the geoip db and that went away 14:21:36 hiro, and how often is the geoipdb updated? 14:21:44 every tor release 14:21:50 especially in relation to this new spike 14:21:54 and then manually for onionoo 14:22:08 raya[m]: maybe we could cross check with psiphon data too - https://psix.ca/d/Vcvqvj6Wk/global-overview?orgId=2 14:22:17 and see if they also saw a spike on their usage 14:22:23 but now that was automated so every two weeks it is updated on onionoo 14:23:47 all the geoIP craziness over the years may mean we should also check against other geoIP sources... 14:23:50 and ooni 14:24:25 i see it all the time just checking manually 14:26:59 sorry.. my point is that if it doesnt make sense, we should have some checks. 14:27:34 but how can we check? we dont have the ips, do we? 14:28:02 we can start based on known IPs... with relays even 14:28:57 but this are the ips of the users that connect to the nodes, how can i check that the country is real, if i dont have that ip? will i need to run a relay for this? 14:29:16 i just meant checking the geoIP data source 14:29:26 sorry if i distracted... 14:29:34 just a long time issue 14:29:45 nigeria used to look the same as uk.. 14:29:53 since it was rolled out as brit telecom 14:30:02 but that was a long time ago :) 14:31:08 alright 14:31:21 raya[m]: sometimes it's very hard to understand what's up 14:31:40 i have been collecting these events here: https://gitlab.torproject.org/tpo/community/support/-/issues 14:32:07 maybe we should add ukraine's invasion to the events timeline? 14:32:14 example: https://gitlab.torproject.org/tpo/community/support/-/issues/40049 14:32:22 emmapeel: david already added 14:32:36 oh great 14:32:41 thank you! 14:33:13 one last thought on ua spike... 14:33:29 in eastern europe, it wasn't uncommon to pick up networks over the border 14:33:37 it actualy is a 'roaming' problem 14:33:41 and we have this monitoring workflow: https://gitlab.torproject.org/tpo/community/team/-/wikis/On-monitoring-metrics-per-country 14:34:01 so possible that those 'ua users' were in 'ru' 14:34:11 due to cell phone shutdowns near the border etc. 14:34:19 just a thought. done. sorry. 14:34:34 gman999: i thought about that! 14:34:51 i had the issue before! 14:34:51 sorry, im laaate! :) 14:37:59 https://gitlab.torproject.org/tpo/community/support/-/issues/40068 14:38:00 hello rhatto! 14:38:32 thanks gus! 14:38:35 np, rhatto. this was a discussion about spike in Tor usage in .ua at end of Feb/ start of March 14:38:40 https://metrics.torproject.org/userstats-relay-country.html?start=2022-02-01&end=2022-03-31&country=ua&events=on 14:40:06 championquizzer, ah, ok! thanks for bringing me back in the loop :) 14:40:28 honestly, i feel like there's a lot of 'parties' who should look at taht collectively.. 14:40:31 including ooni ppl 14:41:49 yeah 14:43:52 i just asked on #ooni 14:44:16 i added some info on the ticket 14:44:21 https://explorer.ooni.org/search?since=2022-02-24&until=2022-02-28&failure=false&probe_cc=UA&only=anomalies 14:45:01 these were the sites that had some issues during the spike period 14:46:49 feel free to add your observations on that ticket 14:46:56 later we can link on metrics timeline 14:47:12 anything else that people would like to talk or share today? 14:47:15 will do! 14:47:34 side note on that... 14:47:43 i have relays 'blocked' from ru in metrics.. 14:47:51 but it looks like i still get ru traffic 14:47:59 but who knows if it's legit 14:48:13 we dont have a mechanism to distinguish good v bad connections 14:49:16 and it worth remember that the metrics data is a conservative estimation of tor usage 14:49:25 https://gitlab.torproject.org/tpo/network-health/metrics/website/-/issues/40009 14:49:33 right. 14:51:16 if everyone is good, i will stop the bot 14:51:36 super 14:51:44 thank you for the feedback 14:51:48 learned a lot 14:51:56 good stuff 14:52:03 thank you for bringing that topic, raya[m]! 14:52:10 thanks all! 14:52:19 Thanks! 14:52:20 yes, ty, raya[m] :) 14:52:32 #endmeeting