How to Test Autoscaling in Kubernetes
In an ideal world, you want to have precisely the capacity to manage the requests of your users, from peak periods to off-peak hours. If you need three servers to attend to all the requests at peak periods and just one server at off-peak hours, running three servers all the time is going to drive up expenses, and running just one server all the time is going to mean that during peak periods, your systems will be overwhelmed and some clients will be denied service.