Yes, multiple disc backups are on the plan, however yes, if like @TonyRogers a filesystem is backed by multiple discs, then the current implementation would not be synchronous across the two discs, so the snapshots would not be much use.
At the moment we're assessing usage, storage network capacity etc to make sure we can accommodate everyone.
Initially we were thinking of using the qemu freeze/thaw commands to let the guest know when a backup was being taken, so filesystem activity could be quiesced. In practice this was an RPC nightmare with complex failure scenarios, mostly risking involving frozen VMs. So we moved to just taking snapshots in as atomic way as we could, relying on filesystems and databases to be able to repair themselves when getting between writes.