16:00:21 <hiro> #startmeeting network-health 08/14/2023
16:00:21 <MeetBot> Meeting started Mon Aug 14 16:00:21 2023 UTC.  The chair is hiro. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:21 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:33 <hiro> ok sorry @rishadbaniya[m] can you tell me your question again?
16:00:56 <hiro> if I miss a question please do ping me and I'll reply to you right away
16:00:56 <rishadbaniya[m]> so in onionperf analysis files...i saw that the within the circuit field there's this path
16:01:04 <rishadbaniya[m]> and that path has 5 relays
16:01:08 <rishadbaniya[m]> and some of them have 1
16:01:20 <rishadbaniya[m]> what's the deal with this different number
16:01:30 <rishadbaniya[m]> i had read in onionperf docs that it doesn't interfere
16:01:42 <hiro> are those onions?
16:01:44 <rishadbaniya[m]> with how circuit is created
16:02:34 <hiro> no it doesn't, it's the same tor process that it would be used anywhere
16:03:17 <hiro> and where do you see only one node, does the circuit succeed? what is the status?
16:03:26 <hiro> and what is the destination?
16:03:40 * rishadbaniya[m] sent a code block: https://matrix.org/_matrix/media/v3/download/matrix.org/LPcOhCnYKytRuIGdEnlREzZP
16:03:54 <rishadbaniya[m]> for example this one has 4 nodes mentioned
16:04:17 <rishadbaniya[m]> hiro: let me search for it again quickly
16:04:49 <rishadbaniya[m]> and what's the deal with "filtered_out", what does it signify?
16:05:03 <rishadbaniya[m]> relay metadata being opted out of the analysis file?
16:07:21 <hiro> it's a filter for the visualization step
16:07:33 <hiro> cbt stands for compute build timeout
16:07:46 <hiro> https://gitlab.torproject.org/tpo/network-health/metrics/onionperf/-/blob/develop/CHANGELOG.md#L33
16:08:26 <hiro> https://gitlab.torproject.org/tpo/network-health/metrics/onionperf/-/blob/develop/onionperf/onionperf#L330
16:08:39 <hiro> sorry circuit build timeout
16:11:02 <rishadbaniya[m]> <hiro> "and where do you see only one..." <- i can only get it once i log it from the parsed data..i wanted to know...there's this case of having only 2 relays in the path, does that mean it extended upto the 2nd hop until an error
16:11:22 <rishadbaniya[m]> example here... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/xtBEXyXuPiwkkBYDPncTNjAM>)
16:12:52 <hiro> I think that's a case where the client tried a circuit, then extended and then the circuit failed... but that's mostly a tor error and I think some tor dev would know more of the details
16:12:58 <hiro> and why that happens
16:14:53 <rishadbaniya[m]> as of right now in the integration part of relay partitioning tool, i'm ignoring the ones where error had occured because of further research on it..
16:15:24 <rishadbaniya[m]> and about the "filtered_out", how should i interpret the path when it's present there
16:16:03 <rishadbaniya[m]> some relay information extracted out? or should i not consider it as something that took out the metadata about the relay from path array
16:16:31 <hiro> I am not sure how that plays out for the analysis you are doing to be honest. I think that's a question for @juga
16:16:50 <hiro> but I think they are traveling to CCC right now and might reply later
16:17:34 <rishadbaniya[m]> Sure :)
16:17:46 <hiro> in my case I tend to filter those measurements but I do not know if the results of those failed paths are acqually useful for what you are doing
16:18:37 <rishadbaniya[m]> so as of right now i'm currenlty using the "successfull" three hop paths to create a two hop  combination out of it A, B, C to A,B and B,C
16:18:50 <rishadbaniya[m]> and using it to minimize the circuit creation attempt
16:20:24 <hiro> yes but that filter is used for guard analysis .... this is the mode that usually triggers that filter: https://gitlab.torproject.org/tpo/network-health/metrics/onionperf/-/blob/develop/onionperf/onionperf#L242
16:20:46 <hiro> I am not sure how that is useful for you to be honest, I am afraid you'd have to wait for @juga to come back online
16:25:48 <hiro> allright if that's all we can end the meeting, but before doing so, please do update the pad
16:25:57 <hiro> https://pad.riseup.net/p/tor-nethealthteam-2023-keep
16:27:38 <hiro> #endmeeting