Are VLANS through the eve-ng cloud truly separated ( at a MAC level )

[kiteboy]

I have a test environment where I'm tunnelling 36 physical links from a Physical Test Environment to a Virtual Environment.
This project has been a fantastic success, allowing incredible test coverage. All the point-to-point links work perfectly, we use QinQ to push all the vlans down one physical link into a dedicated link
in the back of the EVE-NG Bare Metal server.

Where we have run into problems with the Layer 2 Aspects, where there appears to be the potential for the same mac in 2 different vlans. ( this is a common scenario in the real world )
Are the MAC addresses truly separated at the VLAN level?
Are there any commands I can run to troubleshoot or see the MAC-layer in the "cloud" switch? Or any additional documentation to provide a better understanding.

I read the cookbook multiple times, but I couldn't really understand the functionality of QinQ for the Cloud switch. (How do I configure the outer VLAN) I'm using a Juniper QFX in the physical side and a Cisco vIOS switch on the virtual to provide QinQ

Is there a better way to do this. ? I'm starting to consider VxLAN to get myself away from this MAC address issue altogether