Hi there,
It is what we do - have around 150 VMs that we failover regularly. When I say regularly we have a DR exercise every 6 months where we fail production to our secondary site are run there for a week then failback.
As long as the VM is sized correctly don't see any reason why you wouldn't do it - both are required for the exercise so seems to be one less moving piece in the equation. I can imagine if you have a truly massive site then having a separate server for SRM would be good but at our sizes...
Kind regards.