13:00:11 <hiro> #startmeeting network-health 2025-07-07 13:00:11 <MeetBot> Meeting started Mon Jul 7 13:00:11 2025 UTC. The chair is hiro. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:11 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:24 <hiro> and here is the pad 13:00:25 <hiro> https://pad.riseup.net/p/tor-nethealthteam-2025-keep 13:00:51 <sarthikg[mds]> o/ 13:01:10 <GeKo> heyho! 13:01:27 <hiro> I hope everyone has had a good break 13:01:29 <GeKo> everyone survived the scorching heat i hope 13:01:45 <juga> what heat? :P 13:01:50 <GeKo> lol 13:02:06 <GeKo> well, not here. it was max 25°C :p 13:02:45 <juga> tbh, temperatures were even a bit too high for me, 30 inside flat... 13:02:55 <hiro> isn't 25 too hot for where you are? 13:03:47 <hiro> sarthikg: are you think we are complaining for nothing? 13:03:47 <hiro> lol 13:03:52 <sarthikg[mds]> 25 is hot folks? it was upwards of 40 here :( 13:04:06 <GeKo> hiro: almost 13:04:28 <GeKo> :) 13:05:17 <hiro> anyways I have added my updates in the pad, just basically continuing where I left 13:05:25 <juga> same here 13:05:32 <juga> continuing with the db 13:05:39 <hiro> we have a 183 sync next week on monday.. if you haven't been added sarthikg I'll send an invite to you 13:06:26 <GeKo> hiro: there are some loose mrs that require some small changes to push them over the line, i think 13:06:35 <sarthikg[mds]> hiro: I don't have an invite, but I see the meeting in the calendar. the 183 kickoff meeting i believe? 13:06:52 <hiro> yep GeKo (IRC) there are a few of them I'll try to resolve those 13:06:59 <GeKo> thanks 13:07:51 <hiro> sarthikg: yes next monday (14th) at 1400 UTC 13:07:56 <GeKo> hiro: it might be worth thinking about the postgres dump idea so folks could ingest such a dump locally and do some research on their own 13:08:08 <GeKo> not sure if there is much that we need to do here 13:08:20 <GeKo> but it might be cool to have something like that handy next week 13:08:36 <GeKo> i'll have the talk at bornhack then and thought about linking to such a dump 13:08:44 <hiro> how many months you'd like to have? 13:08:49 <GeKo> so interested folks could get their hands dirty 13:09:04 <GeKo> that is one of the questions we have to decide 13:09:16 * juga is curious about what's the postgres dump idea 13:09:21 <GeKo> imo that's mainly a matter of size 13:09:42 <hiro> yes... but also if you want all the tables or some could be excluded 13:09:52 <GeKo> max 1 year, i think, would be a good start 13:09:57 <GeKo> right 13:10:06 <GeKo> let me think a bit about that this week 13:10:43 <GeKo> juga: we thought about linking to some postgres dump over the last n months from somewhere on our wiki etc. 13:11:00 <GeKo> so folks could just ingest that locally and start doing n-h research on their own 13:11:15 <juga> GeKo: the metricsdb dump i guess?, that sounds cool 13:11:15 <hiro> I think 1 year could be a lot of data... I'll do a test for a week and see how that looks like and will try that out 13:11:21 <GeKo> juga: yeah 13:11:25 <juga> i tried that... 13:11:40 <GeKo> aaaaaand? :) 13:11:53 <juga> the issue i had was how to get related tables data, let's say for a week, when you don't know when it was published... 13:12:04 <juga> i started an sql script, but didn't finished... 13:12:21 <juga> (i was focusing on bw data) 13:12:34 <sarthikg[mds]> what tables are we looking at to be included in the dump? 13:12:53 <juga> so, i guess a dump from x to y time, would be usefula for us too 13:13:06 <juga> if we can get correct the related tables dates 13:13:06 <GeKo> that was the idea, yes 13:13:20 <GeKo> we are at the end n-h researchers as well :) 13:13:32 <juga> nice 13:13:35 <hiro> sarthikg: maybe we could exclude families and bridges? 13:13:43 <hiro> GeKo (IRC): do you wanted bridges too? 13:14:08 <GeKo> sarthikg[mds]: i am hesitating to say server_status to begin with 13:14:17 <GeKo> :) 13:14:31 <GeKo> hiro: i think that might not be needed 13:14:52 <GeKo> i think there are two use-cases here. the general one we had in mind for quite some time 13:15:06 <hiro> ok I'll have a look 13:15:06 <GeKo> and the bornhack crowd next week 13:15:26 <GeKo> i'd say we could provide way less for the next week one. 13:15:32 * juga is now tempted to go to bornhack... 13:15:47 <GeKo> maybe just server desc, statuses, consensuses and extra info descs 13:15:55 <GeKo> juga: do it :) 13:16:08 <juga> i'd see :) 13:16:17 <juga> s/'d/ll/ 13:16:27 <GeKo> hiro: maybe try with those 4 above first and see what we get size-wise? 13:18:01 <GeKo> apart from that i'll continue p183 and bad-relay work 13:18:03 <hiro> sure! 13:18:14 <sarthikg[mds]> for the network_status tables, I have a python script that helps me import related data across tables. maybe we can repurpose that.. i'm not sure if its doable with just sql... (https://gitlab.torproject.org/sarthikg/parser-import) 13:18:22 <hiro> oook! anything else? 13:18:31 <hiro> if not we can continue to talk async as usual 13:18:39 <GeKo> sarthikg[mds]: nice, i'll have a look at it 13:18:43 <GeKo> i am good 13:18:43 <hiro> and end the meeting 13:18:51 * hiro is at least for the temperture 13:18:57 <hiro> oook for some reasons I am seeing everybody messages now 13:19:03 * juga is good 13:19:04 <hiro> in random order ooook 13:19:19 <sarthikg[mds]> yeah, irc lagged for a few mins for me 13:19:32 <GeKo> but it seems your keybord issue is solved at lea s t :p 13:19:46 <hiro> lol GeKo (IRC) I changed keyboard 13:19:54 <hiro> gave up at the fancy one 13:20:00 <GeKo> the quick-and-dirty fix, i see 13:20:01 <hiro> I'll blame ggus (IRC) 13:20:07 <GeKo> let's do it 13:20:14 <hiro> anyways! 13:20:19 <hiro> #endmeeting