You are preparing for an upcoming product launch and have set up an autoscaling instance group to handle web traffic. This instance group is configured as a backend service for an HTTP(S) load balancer. However, you observe that the VM instances in the group are being terminated and relaunched every minute. These instances do not have public IP addresses. Despite verifying that each instance can provide the correct web response using the curl command, the issue persists. What should you do to ensure the backend service is configured correctly?