A SysOps Administrator is running an auto-scaled application behind a Classic Load Balancer. Scaling out is triggered when the CPUUtilization instance metric is more than 75% across the Auto Scaling group. The Administrator noticed aggressive scaling out and after discussing with developers, an application memory leak is suspected causing aggressive garbage collection cycle.
How can the Administrator troubleshoot the application without triggering the scaling process?

A. Suspend the scaling process before troubleshooting.
B. Delete the Auto Scaling group and recreate it when troubleshooting is complete.
C. Remove impacted instances from the Classic Load Balancer.
D. Create a scale down trigger when the CPUUtilization instance metric is at 70%.