In general storage definitions are by pool in Proxmox VE. How you setup the pools on Ceph's side is not known by Proxmox VE nor does it need to.
Easiest and most stable solution is to have two different pools.
zcat ceph/ceph.log.4.gz | grep -i skew
Run the above on your logfile. The cluster reports clock skew and that will slow down Ceph considerable. The cluster needs to have the same time, doesn't matter if correct.
Wireguard was included into the kernel in a later version. You will need to build it.
But they do run with the same host kernel. And the wireguard is loaded by the host kernel. Any issue and it will halt the node, not just a VM.
I have it running. Just not in a container. ;) Too much to break for a CT to get it running.
In general, I'd argue that it doesn't make much sense to have a wireguard interface inside a CT. Why do you want it inside a CT?
In a cluster the config lives on the other nodes as well. On a total failure of the node, the configs will be re-synced. Besides the configs outside /etc/pve.
To backup, just copy the whole /etc, whichever tool you use, it needs to work across fileystems (eg. /etc/pve).