When upgrading, did you upgrade both AMP and Failover at the same time? Or did you stagger the upgrade? It's possible that the Failover is behaving as the AMP.
A quick check is to run this CLI command:
# dbc 'select count(*) from ap'
If it outputs 0 then it's still acting as Failover. <-- Failover should always report 0 AP/Devices until it's in Failover state.
If it outputs any number greater than 0, then it's in Failover state acting as the AMP.
A step further would be to make sure the account the Failover is trying to use is still valid on the AMP.
Verify that the AMP can ping the Failover, and vice versa from the CLI.
If all these things are true, then it's possible that the Failover is having problems talking to the AMP over HTTP/HTTPS.
A test for this might be to delete the AMP from being watched and then re-add.
If none of these work, then you'll want to open a support case.