13:01:29 <hiro> #startmeeting network-health 2025-06-23
13:01:29 <MeetBot> Meeting started Mon Jun 23 13:01:29 2025 UTC.  The chair is hiro. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:29 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:02:27 <hiro> how is everyone? I haven't got much for this week except what is on the list on my pad... trying to reduce load on metricsdb-01
13:02:40 <hiro> luckily seems we have sorted out meronense, so one less of a issue
13:03:07 <GeKo> good, good
13:03:18 <hiro> PSA jun 30 - jul 5 all tor is off
13:03:29 <hiro> if you don't remember - I wasn't
13:03:53 <GeKo> yeah
13:04:08 <hiro> hey 4xy0m o/
13:04:10 <hiro> welcome
13:04:16 <GeKo> hiro: q: you closed https://gitlab.torproject.org/tpo/network-health/metrics/descriptorParser/-/merge_requests/107, which was pretty close to be mergeable
13:04:25 <GeKo> do you plan to post a follow-up mr?
13:04:36 <M4xy0m[mds]> hiro: thank you !
13:04:37 <hiro> I didn't? GeKo (IRC) ?
13:04:53 <hiro> sorry that was a mistake yeah I have a fix branch to address your comments
13:05:15 <GeKo> it might be !110
13:05:20 <GeKo> not sure
13:06:04 <GeKo> right, i guess it just waits for getting set to needs_review, okay
13:06:09 <hiro> sorry GeKo (IRC) yes I had https://gitlab.torproject.org/tpo/network-health/metrics/descriptorParser/-/merge_requests/110 ready
13:06:31 <GeKo> alright, no worries
13:07:05 <hiro> ok, who wants to go next with the updates if any?
13:07:33 <GeKo> not much from my side
13:07:58 <GeKo> i did a bunch of p183 work which led me into the materialized views rabbit holw
13:08:00 <GeKo> *hole
13:08:14 <GeKo> and it seems we start digging us out of that one, so that's good :)
13:08:28 <hiro> yeah, some of those were experiments in making the ingestion faster
13:08:39 <hiro> and some I could cleanup already
13:08:59 <hiro> it seems we are writing a lot of data
13:09:20 <hiro> so even though we aren't growing the disks the updates are saturating the link in the cluster
13:09:32 <GeKo> well, most of them run just once a day
13:09:34 <hiro> https://gitlab.torproject.org/tpo/tpa/team/-/issues/42174#note_3212826 has all the juicy bits
13:09:54 <GeKo> i wonder whether the issue tpa has comes from the hourly run of the other ones
13:10:12 <hiro> yeah, but now I have removed all the ones we weren't using
13:10:18 <hiro> so that helped
13:10:35 <hiro> something I was reviewing with juga were the indexes, some got duplicated
13:10:43 <hiro> so we might delete some of those
13:11:05 <GeKo> this started in march, hrm
13:12:15 <hiro> i don't know why march.. the number of relays has not changed
13:13:07 <hiro> so it has to be something w the mat views or indexes
13:14:15 <GeKo> but nothing landed on the begin or march regarding those views, hrm
13:14:22 <GeKo> *at the begin
13:15:30 <hiro> anyways sarthikg juga 4xy0m do you have anything for today's sync
13:16:06 <M4xy0m[mds]> I don't think so
13:16:50 <sarthikg[mds]> not a lot, for me the aggregator is working for relays, doing some final data validations, post which i'll add bridge support as well. i'm also working on adapter for the new tables.
13:17:22 <hiro> oh great sarthikg
13:18:05 <hiro> GeKo (IRC): one thing I noticed is that maybe we have more documents starting in march like we have added between 50MB and 80MB of server descriptors since the previous months (compressed)
13:19:05 <hiro> with consensuses is difficult to notice but the archives have grown...
13:19:24 <hiro> so this + the thing that I have loaded more data into the DB maybe combined did the trick
13:19:36 <GeKo> hrm
13:21:47 <hiro> anyways if everyone is groot we can end the meeting and talk async as usual
13:21:48 * hiro is groot
13:22:02 <sarthikg[mds]> i'm groot!
13:23:36 <hiro> oook! I'll end the meeting
13:23:40 <hiro> #endmeeting