1. 3
  1.  

  2. 2

    If I were to graph the costs as function of requests, serverless (all inclusive) will be an exponential of sort, whereas EC2 etc are more linear/step function. The obviously hard part, is to figure out which one is lower for your business at any time. I’d be curious to see some aws metric/alert magic + autoscaling to shift the traffic between the two architectures depending on load/requests.