What's the device count for both buildings (switches, controllers, APs)?
Following your scenario and expecting that there's no communication issues between Building A and B.
1a) If AirWave in Building A goes offline
- I take it that the AirWave in Building B is a Failover AMP. It will restore the last backup it got from Building A, and then resume collecting data..
1b) Bringing AirWave for Building A back online will result in 2 AMPs monitoring. The option you have while bringing up Building A's AMP is do you care about the data that Building B's AMP caught while Building A was down? If you do, then you take a backup of Building B's AMP and restore it on Building A. Then you restore the failover image on Building B's AMP to make it a Failover AMP again. If you don't care about data that was collected during AMP A's outage, then you can just restore the Failover backup in /var/airwave-backup/watcher/
2a) If all things in Building A go down, Building B's AMP and controllers take over. This means that the controllers become active. It will take longer, but eventually Building B's AMP will show that Building A's devices are down, with the APs migrated to Building B's controllers (APs will remain in same group/folder, but their details will reflect the new controller upstream).
2b) Same as 1b.
The scenario is more difficult to describe if you have AMP A and AMP B as standalone AMPs. As until failure, Building B's AMP would be doing nothing if the APs are all provisioned to Building A's controllers. Unless you monitor both sets of controllers from both AMPs for an active/active AMP setup (but that's not an advised route due to SNMP load and latency).