I’m a computer science student from Germany and next semester I will write my master thesis about pod autoscaling in Kubernetes clusters.
Regarding the topic of the thesis, which I will have to define more precisely, I have a scenario and a question for you:
Imagine you would want to use some kind of docker image from e.g. Docker Hub and you would want to autoscale the resulting pod. Would it be helpful in any way to you to know which general autoscaling strategy (vertical or horizontal) you should use for this pod (to define manual rules, or use as knowledge for a learner)? And why?
Because it makes only sense to scale certain pods in a specific way. It makes no sense to vertically scale a pod that is not able to use multithreading for example…
Please let me know your thoughts, it would be really helpful