15:58:14 <GeKo> #startmeeting network-health 08/21/2023
15:58:14 <MeetBot> Meeting started Mon Aug 21 15:58:14 2023 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:14 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:58:17 <GeKo> alright
15:58:33 <GeKo> welcome everyone to our weekly network-health meeting
15:59:11 <gaba> o/
15:59:21 <GeKo> our pad is, as usual, at https://pad.riseup.net/p/tor-nethealthteam-2023-keep
15:59:55 <GeKo> please add your items if you have not done so yet
16:00:08 <GeKo> juga and ggus will be afk today so might be a short meeting :)
16:00:35 <rishadbaniya[m]> hey there
16:00:40 <rishadbaniya[m]> how are you all :)
16:01:31 <mattrighetti[m]> o/
16:02:07 <GeKo> hi hi!
16:02:21 <GeKo> hiro: how did things work while i've been afk?
16:02:32 <GeKo> i hope nothing exploded?
16:02:35 <hiro> not really
16:02:49 <hiro> except that timestamp error on victoria metrics
16:03:04 <hiro> that was a conversion mistake between timestamps and long
16:03:11 <mattrighetti[m]> Welcome back GeKo :)
16:03:22 <rishadbaniya[m]> been quite long
16:03:27 <hiro> and the website decided to overwrite some of our data... not sure why... glad we are mving away from that giant db'
16:03:41 <hiro> to another giant db lol
16:03:57 <GeKo> heh
16:04:02 <GeKo> okay
16:04:09 <GeKo> thanks for running those meetings
16:04:13 <GeKo> much appreciated
16:04:38 <hiro> no worries
16:04:42 <GeKo> if there is anything urgent anyone of you wants me to prioritize this week while i am cutting down my backlog
16:04:46 <GeKo> please ping me
16:04:53 <hiro> @rishadbaniya[m] you had a few questions last week that I wasn't able to answer
16:04:54 <GeKo> i might get to your thing earlier that way :)
16:05:01 <GeKo> okay
16:05:07 <hiro> were you able to clear those points out?
16:05:14 <GeKo> yeah, we might be able to start with the gsoc parts
16:05:21 <rishadbaniya[m]> yep, it was about onionperf analysis file
16:05:27 <rishadbaniya[m]> i got it fixed now
16:05:32 <rishadbaniya[m]> i've used successful circuit attempts
16:05:38 <rishadbaniya[m]> only stored in db
16:05:42 <rishadbaniya[m]> and assessed
16:06:56 <hiro> ok
16:07:22 <GeKo> rishadbaniya[m]: so, what do you think is left for this week?
16:07:39 <rishadbaniya[m]> juga had added things about prometheus
16:07:41 <rishadbaniya[m]> in the issues
16:07:52 <rishadbaniya[m]> and documentation is left
16:08:41 <GeKo> okay
16:08:59 <GeKo> so, how are you currently determining partitions?
16:09:18 <GeKo> is there a script running over the data collected?
16:10:06 <rishadbaniya[m]> currently there are few errors
16:10:17 <rishadbaniya[m]> that decide the partition
16:10:21 <rishadbaniya[m]> such as no route tho host found
16:11:32 <rishadbaniya[m]> there are other errors too..and these mesages are logged
16:12:13 <GeKo> okay
16:12:15 <GeKo> sounds good
16:12:40 <GeKo> i think i might get up to speed on your latest work by wednesday
16:13:07 <GeKo> is there anything you need from my side or juga's side?
16:13:24 <GeKo> or are you all set on picking up the remaining items and working on them?
16:14:14 <rishadbaniya[m]> yeah i'm
16:14:18 <rishadbaniya[m]> doing 2..3
16:14:21 <rishadbaniya[m]> remaining backlogs
16:14:22 <rishadbaniya[m]> too
16:14:28 <rishadbaniya[m]> and few optimizations
16:14:31 <rishadbaniya[m]> i was waiting
16:14:33 <rishadbaniya[m]> for juga to discuss
16:14:37 <rishadbaniya[m]> about that
16:14:41 <rishadbaniya[m]> he said he is in plane right now
16:14:45 <rishadbaniya[m]> and won't be on meeting
16:15:22 <GeKo> juga goes by they/them fwiw
16:15:25 <GeKo> but yes
16:15:37 <GeKo> hrm, i see https://gitlab.torproject.org/rishadbaniya/erpc/-/issues/10 is still open
16:15:55 <GeKo> so i am a little confused on how you implemented the partitioning analysis
16:16:27 <GeKo> where should i look to get up on this item given that this was the big missing piece when i left?
16:16:42 <rishadbaniya[m]> about that
16:16:43 <GeKo> *get up to speed
16:16:53 <rishadbaniya[m]> juga told me
16:17:48 <rishadbaniya[m]> that we won't be looking
16:17:49 <rishadbaniya[m]> at the errors we can skip them ...we can just log them
16:17:50 <rishadbaniya[m]> and that will be enough for now
16:18:05 <GeKo> i see, okay
16:18:50 <GeKo> rishadbaniya[m]: one thing you should definitely include then, though, is to what we would do if we would look
16:19:04 <rishadbaniya[m]> one of the reasons for this was, the reasoning on the errors that were produced
16:19:04 <rishadbaniya[m]> https://gitlab.torproject.org/rishadbaniya/erpc/-/issues/18
16:19:15 <GeKo> that is what would be the algorithm we'd use for detecting actually partitions
16:19:54 <rishadbaniya[m]> so, the errors will determine the partition, in summary, the errors that we logged into the database
16:20:01 <rishadbaniya[m]> when we try to visualize those errors
16:20:16 <rishadbaniya[m]> by segregating them in specific ways by errors
16:20:32 <GeKo> right
16:20:51 <GeKo> and sounds like a good start
16:21:06 <GeKo> please include that in the docs you probably write this week
16:21:15 <rishadbaniya[m]> Sure
16:21:34 <GeKo> mattrighetti[m]: how is it going for you?
16:21:55 <mattrighetti[m]> Sorry guys, I have a couple of minutes left and then I have to leave so… from my part we have the summary and bandwidths endpoints ready to go, me and Hiro are planning to do some prod tests this week on the prod server to benchmark stuff
16:22:04 <mattrighetti[m]> And I’m working on the vm proxy as I speak
16:22:14 <hiro> yep
16:22:17 <GeKo> nice
16:22:33 <mattrighetti[m]> We still need to make a couple of tests to decide if we want to proxy stuff to vm or the db
16:23:10 <hiro> we should check a little bit the responses times
16:23:27 <GeKo> yeah, makes sense
16:24:21 <mattrighetti[m]> Hiro: let’s get in touch to agree on the plan to deploy the server
16:24:53 <hiro> yep so @mattrighetti[m]  I think I'd try to deploy that on the same vm where the db is running
16:25:13 <hiro> are docs on the repo updated?
16:25:13 <mattrighetti[m]> I agree that would be the ideal scenario
16:27:01 <GeKo> okay, what else should we discuss today apart from the gsoc work?
16:27:09 <GeKo> hiro: i guess we sync on thu?
16:27:14 <hiro> yep
16:27:22 <GeKo> otherwise if there is anything urgent i could do earlier as well
16:27:37 <mattrighetti[m]> I’ll double check the docs but they should be up to date
16:28:10 <hiro> thanks @mattrighetti[m]
16:28:13 <GeKo> i'll focus this week on cutting down my backlog and getting back up to speed on s112 work
16:28:25 <GeKo> that said as a reminder we'll have our bbb sync tomorrow
16:28:31 <GeKo> (so nothing about s112 today)
16:29:01 <GeKo> 1600utc
16:29:43 <GeKo> okay. anything else should discuss today?
16:29:54 <GeKo> or are we all good for now?
16:29:56 * hiro is good
16:30:16 <GeKo> great
16:30:42 <GeKo> then let's keep going with this new week and close this meeting :)
16:30:50 <GeKo> thanks everyone and have a nice week o/
16:30:53 <GeKo> #endmeeting