An incident with our managed database provider has resulted in our primary database cluster entering a bad state. We are currently remediating the issue.

Edit: We've resolved this issue.