How do i setup a Auto scalling group for my ec2 instance where i am running a docker container??

Hello Everyone, I have Ec2 instance of type r6i.large that has 2 vCpu and 16 gb memory in which i am running a fastapi docker container. During the execution, while processing big parts the docker container takes 1 vCpu and 12 gb memory. During this time if any request comes to the fastapi application the request waits for the previous execution to be completed then process it.

Now, i want to setup a architecture where during this execution if a request comes using asg it will scale up another server and process the request there. Then after completion it automatically scale down the server to save the cost. Using asg, i want to scale up another server only when the memory consumption is more than 10 gb and vCpu usage is 1 and another request is waiting in the fastapi application. Can you guys give all the architecture procedure i need to take to achieve this ??