Handle case where all servers in a cluster report different data
It is possible (though very unlikely) that no matter what cluster configuration you use, as long as it requires more than one answer to be identical to another, that it may never get such identical data back.
For example, recently a user told me that they have an address that receives multiple transactions per second and that looking up UTXO is problematic.
We should, at minimum, make sure that we know when all responses have been aquired and reject the request promise if we couldn't reach the integrity threshold.