With the current implemenation the clientInteligence can is set for the complete configuration.

So adding another cluster will use the same approach.

This might be not the best for the 'normal' state, as for the local cluster the client can access every single instance (so HASH_DISTRIBUTION_AWARE), but the backup is only accessible by using a proxy/OCPservice etc so BASIC must be used.

In this case the normal operation must be switched to BASIC as well to be able to failover.

It would be better to set the inteligence per cluster, or perfect if the client is able to detect that automatically.