AWS autoscaling lets you set up groups of EC2 instances which are controlled by a load balancer. The load balancer in turn makes sure your app has the correct number of EC2 instances running at all times. If your traffic is high it adds the maximum that you set. If traffic goes down it adjust to have the minimum EC2 instances that you set.
This system is great for startups who have no idea if their app will go viral or just flop. Often they just flop. But if you are lucky and it takes off you want to be able to handle the traffic so you don’t lose users.
Documentation link to AWS autoscaling
Running EC2 instances at Scale with autoscaling groups – small Ebook that walks through the whole process including using CodeDeploy.