Using AI for autoscaling pods: AI for (almost) every project!

Jolanda Duinhoven

In This session is in English at 19:10 on 23 March 2023 in room Sith

Track: Sith

Recently, I was wondering if we could benefit from some AI in autoscaling pods in Kubernetes. An AI should be able to learn and predict peak and off-peak hours (such as working hours, lunch breaks, end-of-month loads), so we could autoscale up ahead of peaks to get a better user experience and autoscale down on off-peak hours so we save resources. I already found some projects online and heard that colleagues at Alphabet have looked into it as well. I’m going to dive into this and report out my findings in this talk.

Jolanda Duinhoven

Jolanda is a Java consultant at group9, currently working as tech lead at Alphabet. She has a background in Artificial Intelligence and enjoys solving complex puzzles and writing algorithms. Jolanda also takes interest in the organisational side of things, and contributes wherever she can to improve processes and communication. Occasionally, Jolanda sleeps in a tent on top of her car.