16:00:21 <hiro> #startmeeting network-health 2024-09-23
16:00:21 <MeetBot> Meeting started Mon Sep 23 16:00:21 2024 UTC.  The chair is hiro. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:21 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:30 <juga> o/
16:00:49 <hiro> ok first of all welcome @sarthikg ... but he is in a meeting
16:00:56 <GeKo> woo!
16:01:11 <hiro> then the pad: https://pad.riseup.net/p/tor-nethealthteam-2023-keep
16:02:07 <hiro> @ggus since you have to go is there anything from your side?
16:02:50 <ggus> hiro: we should start organizing our agenda for the next meetup.
16:03:27 <hiro> yes! do you know if the otf report is going to be presented?
16:04:19 <ggus> i'm not sure, but it will be public in october
16:05:19 <ggus> (the report)
16:05:30 <hiro> ok I added that to the agenda as I was  wondering if people wanted to discuss that
16:06:51 <hiro> We might have to figure out a few things too as there will be possibly many questions from the community about the press article from last week
16:07:02 <ggus> yess
16:07:30 <GeKo> juga: how hard is https://gitlab.torproject.org/tpo/network-health/team/-/issues/366?
16:07:50 <ggus> i think we should talk about the *health* aspect of it and not exactly about the attack(s)
16:07:51 <hiro> b ut the date is oct 26th right?
16:07:57 <ggus> hiro: correct
16:08:10 <ggus> one week before i have other tor activity
16:08:14 <hiro> ok then we have time to discuss  this when you come back
16:08:34 <juga> GeKo: not sure cause it's a grafana thing, maybe i can manage to overcome that with vm or maybe it isn't fixable, i don't know
16:09:26 <ggus> where can i find the metrics reference on how we calculate guard/exit probability, for example here: https://metrics.torproject.org/rs.html#aggregate/cc
16:09:39 <juga> i'm also finding limits that i'll comment in the issue for the uptime graphs
16:10:10 <GeKo> okay
16:11:02 <hiro> @ggus hold on that is based on consensus weight
16:11:54 <hiro> https://metrics.torproject.org/reproducible-metrics.html#servers
16:12:05 <ggus> hiro: ty!
16:12:45 <GeKo> hiro: i've done https://gitlab.torproject.org/tpo/network-health/analysis/-/issues/67#note_3079577 and thought a bit about that
16:13:16 <GeKo> while the math is not hard for cases where the non-exit weight of exits is 0
16:13:32 <GeKo> we'll need to think about how to count the bytes properly when this is not the case
16:13:40 <hiro> yep I have been thinking about that too... what are your thoughts?
16:13:49 <GeKo> not sure yet :)
16:13:54 <hiro> the current code on the website is possibly broken to estimate onion service traffic
16:14:02 <GeKo> i need to dig a bit deeper
16:14:04 <GeKo> oh?
16:14:11 <hiro> and I was just thinking to go through that for descriptorParser
16:14:11 <GeKo> how so?
16:14:24 <hiro> well not the traffic but I think the number of onions yes
16:14:33 <GeKo> we know it's broken for the number of onions
16:14:34 <hiro> and then I am concerned there is some bug for the traffic too
16:14:42 <GeKo> ah, okay
16:14:58 <hiro> because that is one of the first thing I did in 2021 when I started with metrics
16:15:13 <hiro> and I used step by step an old mr by karsten and some help from asn
16:15:26 <hiro> and I didn't revist it ever since
16:15:35 <hiro> *revisit
16:16:36 <hiro> so my idea was to add this to descriptorParser, to start having onion s tats there
16:17:30 <hiro> and since you have started working on https://bugs.torproject.org/tpo/network-health/analysis/67 we might find some other things that need  to be fixed
16:18:09 <GeKo> could be. it's been a while that folks looked at this kind of stuff...
16:20:33 <hiro> yep
16:21:45 <hiro> GeKo on a side note, next week is p112 meeting right?
16:23:24 <GeKo> yes
16:23:39 <GeKo> 1500 utc on tue oct 1
16:23:45 <hiro> could you invite sarthikg to it?
16:23:59 <GeKo> i'll remind folks on monday
16:24:08 <GeKo> for the tagtor stuff you mean?
16:24:13 <GeKo> otherwise we are almost done
16:24:22 <GeKo> kind of...
16:24:41 <hiro> yep I was thinking it could be helpful, what do you think?
16:25:09 <hiro> the tagtor stuff is usually only the first part
16:25:20 <GeKo> i am fine with that if you feel that hour is worth for him
16:25:34 <GeKo> maybe getting used to sponsor meetings etc.
16:25:47 <GeKo> aka project meetings, heh
16:25:47 <ggus> s/sponsor/project
16:25:52 <ggus> hehe
16:25:53 <GeKo> faster
16:25:55 <GeKo> :)
16:25:56 <hiro> I thought having an idea of how and why we did tagtor
16:26:21 <GeKo> well, yes, but that could be done with the onboarding
16:26:31 <GeKo> i don't think we'll talk much about that in the p112 sync
16:26:35 <hiro> and yeah the meetings too..
16:27:09 <hiro> yeah I went through that... well ok then one more meeting is better maybe since it is already late for him
16:27:17 <hiro> s/more/less
16:27:33 <GeKo> i'd think so but don't have a strong opinion here
16:28:09 <hiro> yeah that's ok, less meetings is better I think
16:29:11 <hiro> ok I do not have anything else from my side... @geko @juga?
16:29:19 <juga> i'm fine
16:29:26 <GeKo> me too
16:29:33 <hiro> ok then!
16:29:36 <ggus> o/
16:29:48 <hiro> is that a bye @ggus? or a raised hand?
16:29:48 <ggus> (saying good bye)
16:29:51 <hiro> haha
16:30:00 <hiro> #endmeeting