Shutting down cluster by setting replicas to 0 does not shut down config listener pod even though there is no reason for it to be up. It only generates connection failure stack traces every second.