Your organization is looking to reduce the overall cost of its Kubernetes cluster while ensuring high availability. The current cluster spans three availability zones and utilizes autoscaling. What is the most effective cost-reduction strategy that still ensures high availability?