We are troubleshooting a failed host in EWR. Apps with a single volume that were provisioned to this host will be unavailable until it's recovered. Apps, and Postgres instances, with multiple volumes should remain accessible for the duration, but with reduced redundancy.