On April 6th catfish.redistogo.com became unresponsive. Typically our alert system would have notified within a minute. But the primary system was self-hosted. Our secondary notification systems do not check as often, but did alert us to the issue.

Once we had been alerted that the server was unresponsive we attempted to reboot the server, but were unable to get the server to reboot. As a result we removed EBS volumes, brought up another instance to restore service to the affected Redis instances.

As a result we are adding more external monitoring services that will alert us when an instance becomes unresponsive.