Hello Curious. Thanks for the feedback. I am not sure about this for databases but you can do something like this with regular web servers or other server with static content used for processing workload. You can achieve this using Auto Scaling Groups. Auto Scaling Groups gives you flexibility to provide a range of instances you want to spin up (or "down") depending on thresholds you set such as CPU usage or memory usage etc. So let's say I start with 2 EC2 instances and specify a maximum of 8, the system will automatically launch new instances to support my workload when CPU exceeds 70% for example and shrink back to 2 during low peak periods.

I have tried the above with Linux web servers running apache with their files sitting on S3. But I think there may be some quirks with servers in an AD Domain (same computer names) and of course database servers that need to keep data consistent.