Discovered while testing an API with the newest version of the SDK, the
health check caused crash loops as Redis wasn't available. We don't want
the service to restart if Redis is missing, as it's an optional service.