ceph on OVH

[PaulVM]

I have a cluster with 2 x SCALE-4 (at RBX), that have 2 SSD for PVE system and 4/6 NVME (Samsung MZQLB). + BK/Storage server
They are using 2 NVME disks as storage so I have 2 spare NVME disks available.
CPU are 2× Intel Xeon Gold 6226R - 16c/32t
RAM 384 GB
Public bandwidth 1 Gbps, Private bandwidth 6 Gbps (at the moment; it can upgrade up to 25 Gbps).

# ping -f internal PING internal 56(84) bytes of data. .^C --- internal ping statistics --- 245376 packets transmitted, 245375 received, 0.000407538% packet loss, time 6710ms rtt min/avg/max/mdev = 0.018/0.020/0.177/0.001 ms, ipg/ewma 0.027/0.021 ms # ping -f extern PING extern 56(84) bytes of data. .^C --- extern ping statistics --- 215873 packets transmitted, 215872 received, 0.000463235% packet loss, time 7920ms rtt min/avg/max/mdev = 0.028/0.030/0.196/0.001 ms, ipg/ewma 0.036/0.030 ms # ping -f bk PING bk 56(84) bytes of data. .^ --- bk ping statistics --- 199504 packets transmitted, 199504 received, 0% packet loss, time 13096ms rtt min/avg/max/mdev = 0.053/0.058/0.187/0.006 ms, ipg/ewma 0.065/0.061 ms

I want to add a third server (also SCALE-4 and also in RBX), and use 2 disks for every server as ceph storage for HA.
In the meantime at least 1 server must be active for currently active VMs in production.

Is this a reasonable solution?
Any advice/hints?
At the moment I have PVE 7.x and I suppose I stay on it (in my experience every major PVE update is non painful on OVH servers at the beginning ..)

Thanks, P.