16:01:25 <GeKo> #startmeeting network-health
16:01:25 <MeetBot> Meeting started Mon May 24 16:01:25 2021 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:25 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:01:29 <GeKo> alright
16:01:57 <irl[m]> hi
16:02:01 <GeKo> hello everyone to a new weekly network-health meeting
16:02:17 <GeKo> ggus: hey, you around today by chance?
16:02:20 <ggus> yess
16:02:26 <GeKo> great
16:02:46 <GeKo> it was not super-urgent last week, so i did not follow up during the week
16:02:56 <GeKo> given that you probably were super busy
16:02:58 <GeKo> so
16:03:19 <dennis_jackson> o/
16:03:31 <GeKo> we still have the [GeKo: ggus will follow-up on that thread] on our pad
16:03:51 <GeKo> we wrote at the first network-health meeting of the new era
16:04:00 <GeKo> i wonder whether that's still a todo thing
16:04:11 <GeKo> or whether you actually already followed-up here
16:04:16 <GeKo> dennis_jackson: welcome!
16:04:26 <ggus> I didn't follow up, but it's on my todo list.
16:04:41 <ggus> i want to review some participants before we push it
16:05:23 <ggus> can you share the meeting pad?
16:05:28 <GeKo> okay, that's cool. i just wanted to make sure it does not fall through the cracks
16:05:31 <GeKo> oh, sorry
16:05:35 <GeKo> http://kfahv6wfkbezjyg4r6mlhpmieydbebr5vkok5r34ya464gqz6c44bnyd.onion/p/tor-nethealthteam-2021.1-keep
16:05:48 <gaba> o/
16:06:16 <ggus> i have 3 things on my plate: 1. that thread, 2. update bad relays page, 3. move the relay expectations to the community portal.
16:06:56 <GeKo> 2. is for the community-portal part?
16:07:02 <ggus> yes
16:07:22 <GeKo> great
16:07:39 <GeKo> i got some relay ops feedback for the bad-relay criteria page
16:07:44 <gaba> gus: im going to add those 3 things to the pad status update for you
16:07:52 <GeKo> and adapted the one we have in network-health land
16:08:04 <ggus> gaba: thanks!
16:08:09 <GeKo> you can see the changes in the git repo if you are curious :)
16:08:18 <ggus> gaba: add this ticket: https://gitlab.torproject.org/tpo/web/community/-/issues/200
16:08:22 <GeKo> they were minor and clarified some things
16:08:49 <ggus> and this: https://gitlab.torproject.org/tpo/community/relays/-/issues/18
16:09:04 <GeKo> fwiw: https://gitlab.torproject.org/tpo/network-health/team/-/wikis/Criteria-for-rejecting-bad-relays is the page we currently have
16:10:07 <GeKo> okay
16:10:45 <GeKo> irl[m]: we were wondering what you think about working on https://gitlab.torproject.org/tpo/metrics/collector/-/issues/40006
16:11:11 <GeKo> do you have an estimate about how much time you would need to fix that?
16:11:23 <irl[m]> i don't know enough about this to understand the risks of exposing those lines so i'm waiting for mikeperry to say something
16:11:30 <GeKo> in a way that general overload* support gets added to descriptor parsing code
16:11:54 <irl[m]> the fix itself is adding the line to a list of lines that are allowed, and then doing deployments of affected code
16:11:54 <GeKo> i think dgoulet is right about that
16:12:02 <irl[m]> the deployments probably take longer than the code change
16:12:29 <GeKo> okay
16:12:50 <GeKo> howe long would it take to add general metrics support for prop#328?
16:12:54 <irl[m]> if we think they are safe, then that's ok i can move forward with that this week to make sure we're not dropping too many descriptors
16:13:10 <irl[m]> what do you mean by general support?
16:13:26 <irl[m]> i'm focussed only on "don't have it break" right now, should there be more to it?
16:13:44 <GeKo> parsing logic for those new overload* lines in metrics code
16:14:16 <irl[m]> so the parsing code will go in metrics-lib at the same time because it's not worth waiting, but having the library parse it doesn't do a lot
16:14:26 <GeKo> well, we eventually need to have support for overload line parsing in metrics code, so we can make better use of those in metrics land
16:14:40 <irl[m]> it's possible that when i'm adding the geoip lookup code in onionoo, i could add a parameter to relays to say they are overloaded
16:15:18 <GeKo> i see
16:15:35 <GeKo> right now i am just concerned with the parsing code being available
16:15:58 <irl[m]> ok so yes, i'll add the parsing code at the same time as fixing bridge sanitisation
16:16:07 <irl[m]> * ok so yes, i'll add the parsing code at the same time as fixing bridge descriptor sanitisation
16:16:26 <GeKo> we could think about the onionoo part in a different ticket i guess
16:16:49 <GeKo> great, so mike is the blocker here then. i can ping him so we get this moved forward
16:16:53 <GeKo> thanks
16:16:57 <irl[m]> sure
16:17:02 <irl[m]> that sounds good
16:17:07 <GeKo> great
16:17:28 <GeKo> do we have anything else to discuss today?
16:17:39 <GeKo> dennis_jackson: how is it going with the experiment?
16:17:44 <GeKo> anything we can help with?
16:18:36 <dennis_jackson> We stopped  the measurement saturday evening. I haven't had a chance to look at it properly yet
16:18:54 <GeKo> nice
16:19:06 <dennis_jackson> but hope to find the time this week to make a quick breakdown and post the data
16:19:28 <dennis_jackson> Hopefully we see either positive impact or at least no impact
16:19:41 <dennis_jackson> Are you planning to look at the overload metrics once the parsing support is in?
16:19:54 <GeKo> yep, re: impact
16:20:58 <dennis_jackson> great
16:21:15 <GeKo> so, we have some non-metrics overload* monitoring ongoing right now
16:21:25 <GeKo> we could look at that
16:21:31 <GeKo> and, yes, that is a good idea
16:21:56 <GeKo> even though this features is not in any tor stable yet
16:22:04 <GeKo> thus, we might not see much
16:22:09 <dennis_jackson> ah true
16:22:13 <GeKo> *feature
16:22:49 <GeKo> i make a note to do so, thanks
16:23:19 <GeKo> anything else for today?
16:23:21 <dennis_jackson> np :-)
16:23:29 <dennis_jackson> nothing from me
16:23:38 <ggus> i'm good
16:23:56 <mikeperry> https://gitlab.torproject.org/tpo/network-health/team/-/issues/44 could maybe use some guidelines/requirements for useful output
16:24:38 <mikeperry> is likely related to https://gitlab.torproject.org/tpo/metrics/analysis/-/issues/33076
16:24:51 <mikeperry> in terms of the kind of tool/script output
16:24:58 <GeKo> yeah, i've talked a bit about it.
16:25:00 <mikeperry> I guess acute likely off today
16:25:07 <mikeperry> ah ok, sorry just dropping in
16:25:17 <GeKo> i am not sure how far acute came with her hours here
16:25:32 <GeKo> no, no, it's actually a good idea :)
16:25:58 <mikeperry> yeah gaba asked me wrt extending her hours to finish that analysis branch.. but we should have an idea of what finished looks like
16:26:11 <GeKo> right
16:26:25 <gaba> yes, we talked about it with geko today
16:26:41 <GeKo> i need to think about that
16:26:53 <gaba> and right, adding more info on what we would be asking her it would be a good idea
16:28:24 <GeKo> mikeperry: while you are here: could you chime in on https://gitlab.torproject.org/tpo/metrics/collector/-/issues/40006?
16:28:47 <GeKo> we could then move forward with that ticket i think
16:31:06 <GeKo> okay, i added the item about those requirements/guidelines to my plate for this week
16:31:23 <GeKo> we'll see how my progress will look like
16:32:00 <GeKo> anything else?
16:32:18 <gaba> not from me
16:32:24 * GeKo wonders whether arma2 will jump in next :)
16:32:24 <irl[m]> not from me
16:32:29 <mikeperry> heh
16:32:48 <GeKo> okay, let's be fast and close this meeting then, thanks everyone
16:32:52 <GeKo> #endmeeting