This message was deleted.

you can use route53 to achieve both create an alb in each cluster and in the route53 create failover records (two identical FQDM), set one as primary and the other as secondary on he active active problem (with GEO location) use route policy of geolocation you can also use the latency policy

Thanks @ambitious-motherboard-40337 for the recommendation. I have created both the kubernetes clusters on on-premise Openstack and VMWare machines. So please share me some on-premise solutions.

OK that is a bit more complex

Thanks @ambitious-motherboard-40337..I will investigate if MetalLB supports that on premise