help with prioritising traffic flow in VLANs between 2 physical datacentres using 2 separate uplinks

Apologies if any of my terminology is slightly wrong, some of this is new to me.

We have recently expanded our customer hosting to utilise a 2nd physical datacentre provided by our colo provider and I'm struggling to route internal traffic between the 2 locations via a new physical uplink they installed. I think I'm fighting against auto-created STP rules that are prioritising one link over the other.

Each datacentre has 100mbit Primary and Secondary internet feeds (for redundancy) and there is a cross-link of 1gbit (again, primary and secondary) that is for internal traffic. Illustration: https://ejquo23388.i.lithium.com/t5/image/serverpage/image-id/23449i6E00D20A296DA84F

I have a Netgear GS724Tv4 at Location A that understands STP and VLANs, and a TP-Link TL-SG1016DE at Location B that understands VLANs but doesn't give me any visibility or control over STP. Screenshot of STP status of netgear switch at Location A: https://ejquo23388.i.lithium.com/t5/image/serverpage/image-id/23450iD6393C377EA5D681/

Ports 17 and 18 are the Primary and Secondary 1Gbit uplinks to location B - Tagged with internal VLAN traffic Port 22 - is Primary 100mb uplink and internet - no VLAN tags Port 23 - connects to another internal switch at location A Port 24 - is Secondary 100mb uplink and internet (redundancy for port 22) - no VLAN tags

Internal traffic (regardless of VLAN tagging) is passing via the 100mb uplinks. If I create a VLAN that excludes Port 22/24 (the 100mbit links) the traffic is isolated to it's own location and doesn't cross over ports 17/18 between location A and B. From what I can tell, it's being dropped by the netgear switch because STP detects ports 17/18 as redundant and is prioritising 22 as the only port to send cross-location traffic through.

My plan is to replace the TP-Link switch with another GS724T and setup Multiple Spanning Trees Protocol (MSTP) to tag internal VLAN traffic in it's own STP that only passes over the 1gb links (ports 17/18) between locations, and doesn't have access to the 100mbit feeds (ports 22/24).

Will this work, or is there another method I should follow to prioritise and route traffic between the 2 locations? What are best practices when routing internal traffic between different physical datacentres? Any advice or hints would be greatly appreciated.