We are investigating an EC2 dedicated host failure affecting a small number of apps and databases. Affected apps are currently being restarted on healthy instances (apps scaled to 2 or more are automatically distributed across availability zones and will automatically failover).