Find an easier way to verify whether a relay consensus weight makes sense

This come from a conversation by IRC with @gk today, in which a relay with less advertised bandwidth can have more consensus weight than other with more advertised bandwidth (by 20MiB).

Having the ratio (sbws#40156 (closed)) would help a bit, though we'd still need the previous consensus for the measured relays and the advertised bandwidth for the unmeasured.

Creating this issue here, cause we should think about whether this implies modifications to sbws, onbasca, bwauthealth and/or metrics.