Something like has actually happened before in 2012 where a portion of ELB state data was logically deleted and customers were not able to manage their load balancers. Funny thing, they were able to create new load balancer but not able to manage the existing ones. The entire article explains the event in detail.
But in such situations, there are three possible ways to get out of it:
- Wait for AWS to restore service
- Route traffic directly to instance public IPs or elastic IPs if available
- Move your app and traffic to a different region or different provider