

I’d rather try the mail contact option, should work fine if you create them via PowerShell.
Regarding the Internal Relay option: If you configure a Domain with that option then Exchange Online will accept mails for all addresses with that domain and not send any Non Delivery Reports, because it knows that there are more email systems with that domain. EXO doesn’t have the full picture so it has to accept all mails. If there is no internal recipient defined with that mail address the message will get forwarded to an external system that you configured. That 2nd system would need to send the NDR then. Please note: But that configuration would potentially forward more emails than you would want to the external system.

You mentioned that you can’t ping the remote VM. Have you checked if the remote VMs IP is reachable? Is it in the ARP table? Check with
arp -a