I have configured Auto Scalar rules with Container Memory threshold min 30% and max 70%. When the application is getting used the requests are going to instances in round robin fashion. So when the first instance is reaching its threshold but auto scalar will not create new instance ; because it create new instance based on the average.
As #1 instance, reached its limit ; the instance is crashing and the requests to that instance are failing .
Can we configure rules at instance level also?