I have a rails app running on an AWS autoscaling group with a warm pool of 2 stopped instances. When I deploy my code, I deploy code on the running instances, then I create a new ami that will be used for the future instances that will scale-out. My problem is: the code is never deployed on the warm pool instances, because I cannot deploy on stopped instances (I think)... How could I solve this issue?
I consider using lambda function to terminate warm pool instances when a deployment occurs.