You administer an Always On availability group hosted on two SQL Server on Azure VMs deployed across two Azure regions, placing the replicas in different subnets. The availability group listener has two IP addresses, one in each subnet. After a failover to the secondary region, a .NET application experiences long connection delays (roughly 20 seconds) before it eventually connects to the listener. You must minimize these connection times without changing the network topology. What should you do?