I’m going over the documentation for autoscaling v2 apps and I had a question in regards to setting both auto_start_machines and auto_stop_machines to true.
I’m wondering if it is possible to achieve autoscaling within a fixed upper/lower bound of machine counts. Ideally I’d like to always have at least one machine for my app running to avoid cold start penalties but still be able to scale down after recieving enough traffic that auto_start_machines has scaled up my machine count. I’d also like to be able to set a max number of machines that my app can scale up to, in order to make sure I avoid a surprise bill.
So to simplify my question, is there a way to configure a min/max count when setting
auto_stop_machines = true and
auto_start_machines = true?
It seems like my closest option to achieving what I want if this functionality doesn’t exist is to set
auto_stop_machines = false and
auto_start_machines = true and then set my scale count to 1. That way I have one instance running but can still scale up to meet demand. But with this set up I can’t automatically scale back down and also could potentially scale to an undesirably high amount of machines?
Let me know if this question needs more clarity!