Setting CPU and memory allocations on Kubernetes is a science, not an art. Unfortunately, open-source tooling has been historically lacking, causing engineers to fall back to haphazard guesses and heuristics, rather than precise measurements. Not anymore.
In this talk you will learn:
- Where traditional tooling like the Kubernetes Vertical Pod Autoscaler falls short
- How to determine requests and limits using the next generation of tools
- How to monitor and maintain your resource allocation over time.
By the end of this talk, you will have the confidence to scan your own clusters and take action. You will be ready to optimize utilization, cost, and reliability in your own environments.