You are operating an application hosted on multiple virtual machines (VMs) within a managed instance group that utilizes autoscaling. The current autoscaling policy ensures that if the CPU utilization on any instance exceeds 80%, additional VMs are added to the group. This scaling up continues until the group achieves its maximum size of five VMs or until CPU utilization falls back to 80%. The initial delay for HTTP health checks on these instances is configured to 30 seconds. However, each VM requires approximately three minutes to become fully operational and available to users. It has been observed that during the autoscaling process, the instance group tends to provision more VMs than necessary to adequately handle the end-user traffic. Your objective is to accurately maintain the size of the instance group when autoscaling. What adjustments should you make?