13:01:11 <hiro> #startmeeting network-health 2026-03-02
13:01:11 <MeetBot> Meeting started Mon Mar  2 13:01:11 2026 UTC.  The chair is hiro. Information about MeetBot at https://wiki.debian.org/MeetBot.
13:01:11 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:01:13 <GeKo> awesome!
13:01:18 <GeKo> i can go first
13:01:24 <GeKo> my report is brief enough
13:01:27 <hiro> yuhu!! 🎉 you should go out and celebrate]
13:01:38 <GeKo> i was spending most of my time reviewing branches
13:01:46 <GeKo> above all aggregator!5
13:01:58 <Rohithh[mds]> :)
13:02:03 <GeKo> and it seems this week won't be much different :)
13:02:08 <GeKo> that's all from me
13:02:33 <hiro> thanks GeKo (IRC) ! me and sarthikg will sync tomorrow and we can go through all the needed changes
13:02:44 <GeKo> sure
13:02:52 <sarthikg[mds]> o/
13:03:11 <GeKo> sarthikg[mds]: i wonder whether it would be smart to fix up aggregator!5 first
13:03:24 <GeKo> and then rebase the bridge related stuff on top of it
13:03:43 <GeKo> or should i do aggregator!6 while we work on the fix ups for aggregator!5?
13:04:22 <sarthikg[mds]> GeKo: i was thinking of doing the same. since the bridges uses a lot of logic written already for relays, it makes sense to first get done with relay mr
13:05:43 <sarthikg[mds]> GeKo: i think an initial review to see if there are some data issues, like missing/incorrect data would be helpful. bcoz i have seen some parallels between relays and bridges when fixing or adding data
13:06:33 <GeKo> okay
13:07:11 <GeKo> i'll check that for !6 and do a final thorough review once the bridge changes are rebased on top of the relay ones
13:07:13 <sarthikg[mds]> also my update: i am working on minor updates to aggregator as requested, and i am parallely working on writing a query builder similar to sea-query for clickhouse
13:08:11 <hiro> thanks both!
13:08:54 <sarthikg[mds]> [@juga:matrix.org](https://matrix.to/#/@juga:matrix.org), [@hiro7:matrix.org](https://matrix.to/#/@hiro7:matrix.org): just wanted to check once if any of you came across a query builder for clickhouse in rust. i couldn't find one...
13:09:39 <hiro> no I haven't... is a reimplementation of sea-query a lot of work? but I guess the manual query building would be too
13:10:36 <sarthikg[mds]> hiro: yeah, it seems like it would be quite some work. adding a lot of filters without a query builder is the only alternative, which i am not very inclined to do
13:10:45 <hiro> uhm
13:11:45 <hiro> would you need this for inserting the data mainly or querying the db?
13:12:11 <sarthikg[mds]> hiro: this would be needed in nsa to query the data
13:12:21 <hiro> ah for nsa
13:12:22 <hiro> I was about to suggest to retrieve full tables and do the filtering in memory
13:12:56 <hiro> and what about using the http endpoint in clickhouse and do the query in json?
13:13:02 <sarthikg[mds]> hiro: i was thinking of making use of a lot of prewhere's, as they boost the query performance by a considerable time
13:13:15 <hiro> ah I see
13:13:28 <hiro> uhm
13:14:15 <hiro> ok let's talk about it tomorrow in our sync and brainstorm something. the other thing to consider is that, the library doesn't have to cover all the corner cases but mainly the queries we need and we could also release that which could be nice
13:14:31 <hiro> especially if people start picking that up and help w maintenance
13:15:26 <sarthikg[mds]> hiro: yess! i was thinking the same. i bet there would be a lot of devs using it, since everyone would like to migrate from raw string queries. if its there, i would want to use it in aggregator also
13:15:39 <sarthikg[mds]> let's discuss on this tmrw in our 1:1
13:15:41 <hiro> nice!
13:15:54 <hiro> @geko @juga what do you think of https://gitlab.torproject.org/tpo/network-health/sbws/-/issues/40248  ?
13:17:17 <Rohithh[mds]> sarthikg: sir can you just update me what would be the method of writing a query after you get decided?
13:17:37 <GeKo> hiro: i am fine doing that in case that's what roger really wants
13:17:48 <Rohithh[mds]> I just wrote raw sql string for now for bandwidth route
13:17:49 <sarthikg[mds]> Rohithh[mds]: yes yes, i will!
13:18:02 <GeKo> i wonder, though, whether it would be enough what micah said
13:18:15 <hiro> GeKo (IRC): yes that is what I was thinking
13:18:22 <Rohithh[mds]> Thank you sir sarthikg
13:18:54 <hiro> ok I'll put that in the backlog for now
13:19:36 <hiro> on my side I am stabilizing collector... I hope it will get better now
13:20:09 <hiro> I am also fixing a few of pending stuff for the website
13:20:53 <hiro> the timeline I think has something that is preventing it to be shown for newer events
13:20:54 <hiro> so I'll pick that up again
13:21:12 <hiro> and I'll finish the testnet setup for arti relays
13:21:21 <hiro> that's all for me
13:21:23 <juga> hey, sry my delay
13:22:16 <hiro> no worries! juga want to go next?
13:23:04 <juga> ok, i finished a version of the anomalies implementation
13:23:46 <juga> this week will continue with collector-rs, anomalies fixes and some short issues opened for some time
13:23:49 <juga> that's from me
13:25:00 * juga is going to update pad in abit
13:25:01 <hiro> ok thanks!
13:25:05 <Rohithh[mds]> not much from my side just getting back onto nsa
13:25:16 <hiro> thanks Rohithh !
13:26:39 <hiro> ok if everyone is groot we can go back to chat async
13:26:43 * hiro is groot
13:26:50 * juga +1
13:26:51 <Rohithh[mds]> I'm groot
13:27:22 <hiro> alright!
13:27:26 <hiro> #endmeeting