An application is running on Amazon EC2 instances behind a Classic Load Balancer. The instances run in an Auto Scaling group across multiple Availability Zones. Occasionally multiple incoming requests will receive a 5xx HTTP response when making a request to the Classic Load Balancer. From the Amazon CloudWatch metrics, a SysOps Administrator observes the Elastic Load Balancing (ELB) SpillOverCount metric to be greater than zero during these occasions. These errors can be avoided by triggering scaling actions on which ELB metric?