Amazon EC2 Auto Scaling publishes group-level metrics, including the number of in-service instances. A CloudWatch alarm can evaluate whether that count remains above 180 for a continuous 60-minute period. A step scaling action can then reduce desired capacity, allowing the group to support short-lived surges up to 300 instances while limiting the cost of sustained excess capacity. Scheduled scaling cannot handle unpredictable surges, and maximum instance lifetime replaces individual instances rather than controlling total group capacity. Directly terminating instances through an alarm would bypass normal Auto Scaling capacity management. The original option’s wording is technically imperfect because step scaling adjusts desired capacity; it does not directly rewrite the group’s maximum-size property. Option D is correct when interpreted as a scale-in action that reduces desired capacity to approximately 180 instances.

================