EC2 Auto scaling

You bring up a good questionabout scaling AWS infrastructure to meet changing demands. There are two main approaches here - vertical scaling (scaling up/down an individual instance) and horizontal scaling (adding more instances).

For your single Windows Server 2022 instance, you cannot automatically scale the vCPUs up and down on the fly. With EC2 instances, you would need to shutdown the instance, create an EBS snapshot to save the data, launch a new instance with more vCPUs, and attach the EBS volume. This process allows you to "scale up" that instance, but requires brief downtime.

Auto Scaling allows you to define rules to automatically launch more EC2 instances (horizontal scaling) when certain conditions are met (e.g. CPU utilization).

The key difference is that Auto Scaling and load balancers handle horizontal scaling across servers, while vertical scaling requires stopping, resizing, and restarting a single server. For most workloads, I'd recommend architecting for horizontal scalability to start. But for your use case verticle seems to make sense. Also be aware that you can stop the instance when you are not using it to save your budget.