Compare bandwidth values across relays

Looking at https://consensus-health.torproject.org/#relayinfo the bandwidth values for a concrete relay, they are quite different between torflow bwauths.

This might be due the relays getting stuck in some partition, the network of the bwauth, the location of the bwauths or the location of their Web servers.

For the one relay i looked at, longclaw's value was actually close to the median, which i think makes sense if the reason for the differences between the torflow bwauths is the partitions.

We might no need to compare this if the Total consensus weights across bandwidth authorities is similar between torflow and sbws.

In any case, this would be less time than implementing #40040 (moved).