Fail-safe entry point balancer?

Question

High availability

Actually there is an entry point http traffic ( http://domain.com), which is resolved through haproxy.
There is a task to reserve this entry point as much as possible. That is, imagine the situation, the entry point located in the DC falls off (an accident in the DC), and the services behind it in another (DC) work, but there is no traffic, since the entry point lies somewhere.
Naturally, the question of reserving the entrance arises. But only Round Robin DNS comes to mind. And something tells me that this is never the right decision.
What do you advise?

Answer 1

You probably need to look towards the balancers?

Answer 2

Well, it’s standard to automatically transfer via dns to another DC, and here roundrobin is not very clear.