You are deploying a new service to production that requires automatic scaling via a managed instance group and multi-region deployment. Each instance demands significant resources, necessitating capacity planning. What is the recommended approach?