The standby IP is only used with clustering, so the 0.0.0.0 is expected.
Do your controllers have public reachable IPs?
Or can the controllers be reached by NAT?
Do you have a unique IP/nat per controller?
How do you point your RAPs to those IPs?
If you have a single public IP, the use of VRRP can help here if you point the NAT to the VRRP. If it points to a single controller IP, if that controller becomes unavailable there is no way for the RAP to reach the controller.
If you have multiple public IPs, you can either use LMS/Backup-LMS as suggested before or use round-robin DNS by pointing the RAP to a DNS name that has both (or all) public IPs of your controllers. When using LMS, make sure you point to the public IPs.