DRBD does allow for stacking of resources which could replicate to a third node, however, it would require that the third node be a Linux O/S with DRBD with access to a raw partition. In your case, the third node is a NAS which could not act in this capacity unless you somehow tethered it to another Linux host.
Another consideration is that DRBD will blindly copy all data to all nodes which does not protect you from data corruption.
Since the third SR is just a NAS, I would consider storing entire VM backups there instead which gives you the added benefit of storing multiple backup images per VM. There are many options for doing this in an automated way. There is also one bundled with halizard (however it is not documented). You can get more details by typing "/etc/ha-lizard/scripts/vm_backup.sh --help" from the shell of any host that has halizard installed.