Hello,

for about 3 months I have been hosting two Tor relays:

https://metrics.torproject.org/rs.html#details/AF42E6C77196A37F041A1A1E953E51B4656BDC1B

https://metrics.torproject.org/rs.html#details/7F7D1A5BE88FA7C9358955705AE7AFA61EEDA2B0

After doing system maintenance (mainly upgrading packages), slightly adjusting my "BandwidthRate" and "BandwidthBurst" for both relays and then rebooting the server earlier, the guard relay now shows the following on Metrics, even though the graphs still show the old, correct data:

The first time that this relay was seen in the consensus: 2024-01-10 16:00:00 (8 hours 57 minutes and 21 seconds)

This is even though metrics shows an uptime of almost 16 hours?!

The fingerprints did not change according to both the metrics page, and the Tor log.

I did not delete or mess with each relays DataDirectory or the secret keys.

I have a MyFamily setting in both relays configs:

MyFamily $AF42E6C77196A37F041A1A1E953E51B4656BDC1B, $7F7D1A5BE88FA7C9358955705AE7AFA61EEDA2B0

Yet the relays don't show up as being in a family, and Metrics doesn't at all display the other relay counterpart as a family member when visiting the page for the guard or exit.

There is also no line in the relays journal (from journalctl -u tor@guard) indicating that the guard relay is new to the network, and it retains its older flags.

This only seems to affect my guard relay, the exit metrics page shows mostly valid data, except for the missing family member.

Since I did not change anything that would cause this, this is likely a bug with the metrics page, and I would encourage you to look into it.

Thank you very much,
George