13:00:43 <hiro> #startmeeting network-health 2025-10-06
13:00:43 <MeetBot> Meeting started Mon Oct  6 13:00:43 2025 UTC.  The chair is hiro. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:43 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:01:08 <hiro> https://pad.riseup.net/p/tor-nethealthteam-2025-keep
13:01:23 <GeKo> o/
13:01:53 <GeKo> i am on a train so, we'll see how that goes with the internet...
13:03:07 <hiro> no worries GeKo (IRC)
13:03:46 <sarthikg[mds]> my update: created migration scripts in rust for importing some data from clickhouse locally in json columns to test the compatibility & performance with rust. let me know if anyone else is experimenting with the same, i can separate out the logic and put in a diff repo. also, finalising on a similar setup for clickhouse migrations like that of postgres with datamodel. hiro: just saw the datastore repo. are we planning to have
13:03:46 <sarthikg[mds]> clickhouse & psql models in separate crates?
13:04:20 <sarthikg[mds]> also, clickhouse queries work well with rust, the perf seems good too. there are some datatype gotchas, but apart from that, all looks good
13:04:52 <hiro> sarthikg: we said we wanted to keep them separated right? because there aren't migrations in clickhouse like in postgresql
13:05:39 <hiro> also I have the setup working and can give you credentials to try with metricsdb-03 tomorrow at our sync
13:05:58 <sarthikg[mds]> yeah, we won't need migrations for clickhouse, yes! but i was thinking of keeping the table definitions somewhere so if someone wants to spin-up a local clickhouse for development, it should not be very tough
13:06:21 <sarthikg[mds]> hiro: ohh already? great!!
13:06:48 <hiro> ok let's see, I think the datastore could do that easily? there are some views in there to create the stream tables that might be more complicated to port in code
13:06:52 <hiro> but let's see
13:07:30 <hiro> I was also wondering since there are a lot of code that changed in descriptorParser how I should prepare the MR for this, any idea GeKo (IRC) juga sarthikg ?
13:08:24 <juga> split in diferent commits dedpendnig on functionality?
13:08:26 <GeKo> is it possible to do it in some form of contained commits
13:08:33 <GeKo> yeah
13:09:06 <GeKo> it's fine imo if things break within a set of commits in an mr
13:09:21 <GeKo> if it makes it easier to review the whole thing
13:09:31 <hiro> ok sounds good, should I split the different parsers then?
13:09:37 <GeKo> wfm
13:09:57 <GeKo> might be enough of scoping
13:10:04 <juga> wfm too
13:10:08 <hiro> ok perfect thanks
13:11:11 <hiro> ok that is all from me
13:11:19 <juga> my plans for this week: sbws release and maybe continue with collector, depending on time
13:11:49 <juga> (that's all from me)
13:12:03 <GeKo> i'll do p112 reporting
13:12:22 <juga> and review my mrs :P
13:12:23 <GeKo> hiro: lemme know whether you want some help with the p183 one
13:13:04 <GeKo> a bunch of mrs will be on the plate
13:13:04 <hiro> GeKo (IRC): we have a month on that I am going to have a look today at the document but thanks I'll certainly ask you for a review
13:13:29 <GeKo> then i finally want to get back to the anomaly research part
13:13:44 <GeKo> ponder analysis#80 more
13:13:44 <tor> Uhm, which one of [tpo/network-health/analysis, tpo/network-health/metrics/analysis] did you mean?
13:14:10 <hiro> nice
13:14:13 <GeKo> and there is some follow-ups i need to do for the meetup we had (but that's spare time stuff anyway)
13:14:28 <GeKo> oh, and i need to read up more on clickhourse it seems
13:14:35 <GeKo> *clickhouse
13:14:38 * juga too
13:14:38 <hiro> also we got a request of a M.Sc. student that wants to do a thesis on net health and is looking for something to work on
13:14:47 <GeKo> that's all
13:14:48 <GeKo> yeah
13:15:15 <hiro> so if you have any idea of a problem that could fit that please let me know
13:15:32 <GeKo> will do
13:15:38 <hiro> the student is looking for something a little more applied than research like
13:16:00 * juga thinks on the erpc partitions... but might be small work
13:16:41 <hiro> yeah I was wondering what was left there to do and that could fit a thesis
13:17:06 <juga> not much, i think, so maybe not enough for a thesis
13:17:59 <hiro> ok that's all
13:18:02 <juga> GeKo: if you don't have time for the reviews this week, i can also continue with collector, so no problem
13:18:37 <GeKo> no, will do them tomorrow
13:18:41 <GeKo> all is good
13:18:43 <juga> ok, thanks
13:19:05 <GeKo> or even today :)
13:19:19 <juga> :)
13:19:42 <hiro> ook! if you are all groot, we can chat async
13:19:50 * hiro is groot
13:19:53 <GeKo> o/
13:19:56 <sarthikg[mds]> o/
13:20:03 <juga> o/
13:20:08 <hiro> #endmeeting