network-health

network-health@lists.torproject.org

June 2026

3 participants
2 discussions

Mid-month descriptors from collector-rs?
by David Fifield 04 Jun '26

04 Jun '26

It used to be that the "archive" directory of collector.torproject.org would serve unchanging descriptor tarballs from past months, along with a more or less continuously updated tarball for the current month. But now it seems that there is no tarball for the current month. See here: https://collector.torproject.org/archive/bridge-descriptors/extra-infos/ https://web.archive.org/web/20260504132617/https://collector.torproject.org… The date is currently 2026-05-04, but the latest tarball is extra-infos-2026-04.tar.xz 2026-05-02 03:42 218M Is there a way to get, for example, descriptors from May 3, without waiting until the end of the month? There is the "recent" directory which might work, but it only goes back a few days. At this moment, "recent" covers from 2026-04-30 to 2026-05-04, but if I were to check on 2026-05-20, say, I might not be able to get descriptors from 2026-05-10. https://collector.torproject.org/recent/bridge-descriptors/extra-infos/ https://web.archive.org/web/20260504132701/https://collector.torproject.org… The reason I'm asking: When I make monthly graphs of snowflake bridges, I usually wait a few days into the next month and download the next month's descriptor to include it with the data. The reason is that the tarball for May, for example, contains descriptors that cover the end of April. Without including the beginning of the next month's data, I found, the last 1–2 days of the previous month's data can be incomplete. For example, I just used https://gitlab.torproject.org/dcf/snowflake-graphs to compute statistics using tarballs up to extra-infos-2026-04.tar.xz. This is showing just one bridge (which consists of 12 tor instances). Notice the last few days of April 2026, the "coverage" drops from from 12.0/12.0 to 9.95/12.0: date, fingerprint, transport, users, num_instances,coverage 2026-04-24,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14592.87,12.00,12.00 2026-04-25,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14615.31,12.00,12.00 2026-04-26,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14992.47,12.00,12.00 2026-04-27,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15056.68,12.00,12.00 2026-04-28,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15250.09,12.00,12.00 2026-04-29,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15034.07,12.00,11.92 2026-04-30,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,12877.20,12.00,9.95 If I regenerate the statistics, manually including also the "recent" descriptors, the coverage gets "filled in" (and the users count correspondingly goes up): date, fingerprint, transport, users, num_instances,coverage 2026-04-24,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14592.87,12.00,12.00 2026-04-25,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14615.31,12.00,12.00 2026-04-26,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14992.47,12.00,12.00 2026-04-27,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15056.68,12.00,12.00 2026-04-28,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15250.09,12.00,12.00 2026-04-29,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,15099.58,12.00,12.00 2026-04-30,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14834.64,12.00,12.00 2026-05-01,5481936581E23D2D178105D44DB6915AB06BFB7F,snowflake,14781.53,12.00,12.00

2 2

April and May Microdescriptor Consensuses?
by Ellis Fenske 02 Jun '26

02 Jun '26

Hello, I am an academic working on some Tor research and am trying to download the microdescriptor consensuses and microdescriptors for May and April, and it seems on the CollecTor website that these consensuses are missing from the archive (along with many microdescriptors for April). The regular descriptors seem to be there for May, and recent microdescriptors consensuses are available (though the router status entries in them have future-dated timestamps), but I don't see a way to access the microdescriptor information for this time window. I was hoping you could let me know whether to expect this data to be filled in later, whether I'm misunderstanding something about the way the data is stored, or whether it's just missing. Thanks so much! -- -Ellis

2 2