An entertainment company has launched a new game and initially deployed 12 r6g.16xlarge memory-optimized Amazon EC2 instances behind a Network Load Balancer to ensure player experience during the launch period. The operations team utilized the Amazon CloudWatch agent with a custom metric to monitor memory utilization. Post-launch analysis revealed that the instances were consuming only about one quarter of the expected CPU and memory resources. As demand has become more variable, the company now aims to configure an Auto Scaling group that dynamically scales based on CPU and memory usage to meet fluctuating demand in the most cost-effective manner. What configuration should the solutions architect implement for the Auto Scaling group to achieve this goal?