Hello and welcome to the Horizontal Pod Autoscaling issue + Kubectl productivity!
This might actually be Horizontal Pod Autoscaling Issue part one because when we looked there were over 41 articles that were super cool to include. We put together articles that went beyond the standard fare of setting it up and using it from Control theory to Auto Scaling on Memory with JVM containers. Enjoy!
Wow, this is much more than just a simple article of how to type less - its really the most comprehensive overview on best practices of kubectlrey and a deep dive on how kubectl works. Highly recommended.
Control Theory applied to Kubernetes as it should be. Unlike other “controllers” the autoscalers are much more like classic feedback controllers in that their inputs and outputs are continuous values rather than heuristics-based decisions. Read (and watch) this to find out more.
“moderately in the morning, a little bit during the day, and a lot in the evening” – A classic use-case for horizontal scaling. This is a great recent article on a typical autoscaling journey we see from CPU HPA -> how does it work with cluster autoscaling -> using custom metrics.
Now you can schedule based on metrics collected by Istio! If nothing else, this is a wonderful use case and example for the kube-metrics-adapter from Zalando.
What’s fantastic about this article is the fact that its horizontal pod autoscaling on memory with java so you get a little taste of what it looks like to autoscale on memory while using the JVM.
A horizontal autoscaler operator based on annotations on the deployment. It’s so neat and elegant. Marking autoscaling config on a deployment like this seems to make a bunch of sense.
Ask and you shall receive.