Auto Scaling (ASG) is the 'Elastic' part of EC2. It ensures you have just enough servers to handle the load—no more, no less.
Dynamic Scaling: Scale based on a metric (e.g., 'Add a server if average CPU > 70%').
Scheduled Scaling: Scale based on time (e.g., 'Add 10 servers every Friday at 5 PM for the weekend rush').
Predictive Scaling: AWS uses ML to look at your traffic patterns and scales ahead of time.
ASG doesn't just add servers; it replaces 'Dead' ones. If your .NET API stops responding to the Load Balancer's health check, ASG will automatically terminate it and spin up a fresh one. This is 'Self-Healing' architecture.
Q: "What is the biggest risk with Auto Scaling?"
Architect Answer: "The **Infinite Bill**. A bug in your code (like a busy loop) could trigger a scaling event that spawns 100 servers you don't need. Always set a **Max Capacity** on your ASG and configure **CloudWatch Billing Alarms** to catch spikes before they become expensive."