Move to Standard LoadBalancer Internally
I would like to suggest that you move the platform off of Basic Azure Load-Balancer and instead use the newer Standard ALB. This will open up additional options to us such as use of Availability Zones, Geo-Replication over Peers (vs having to setup more expensive and complicated vNet to vNet VPNs). I'm not sure why you opted for the older Basic ALB over Standard (what I got back was due to cost) - but that doesn't make sense to me as Azure SQL MI isn't exactly a cheap solution anyway and most large customers who would consume it, would prefer the additional features.
John Lithgow commented
+1 for upvote on this issue also.
+1 for this request,. Unable to access Managed Instance from Global Peering VNets.
+1 for this request!
I am unable to consume SQLMI across a Globally peered vNet due to the internal basic load balancer. This is a considerable problem in a hub and spoke VDC design, where the hub is in a different region to MI. This constraint is not apparent in standard load balancers.
Please consider migrating to the standard load balancers, or even giving the option of one.
Please do upgrade we need to move roll back our entire express route change because of this. this is pathetic, why cant you upgrade the same.
This is a very relevant feature. vNet2vNet tunnel is not optimum to have access to MI on a different region. It is a big hassle to be in Region A and not be able to query database in region B. global peering is the perfect answer. However the Basic Load Balancer limits that implementation. please upgrade.