Make Guard flag easier to get
Currently if you restart relay every few days its difficult to get Guard flag. Guard should have continuous availability, not long uptime.
I propose to change measurements for guard - do not use uptime for calculation, calculate only downtime.
This will allow more relays to get Guard flag and shift guard flag and workload from exit nodes to new location. There are only 1300 guards but 2500 relays with stable.
Trac:
Username: hsn