Resource Requests and Limits Under the Hood - The Journey of a Pod Spec

Explore the inner workings of resource requests and limits in Kubernetes through a visually engaging conference talk. Dive deep into the journey of a pod spec, from setting restrictions to their implementation at the Linux kernel level. Learn how Kubernetes manages compute resources through fun hand-drawn illustrations, gaining insights into the transformation of application deployment from traditional methods to container orchestration. Understand the intricacies of resource management, including QoS classes, CFS quotas, and the differences between CRI and OCI runtimes. Discover the process of pod autoscaling and the role of Vertical Pod Autoscaler (VPA) in optimizing resource allocation. Gain a comprehensive understanding of Kubernetes resource management to enhance your ability to efficiently deploy and manage containerized applications.

Intro
Doggy Daycare Analogy
Here's an ordinary pod spec
Requests for planning
Update Cluster State
Scheduler
Create Pod on Node
The Journey of a Request
Requests passing through
Requests summary
QoS Class in Kubernetes
Now let's talk about limits
Limits enforce.
Assign Pod to Node
Kubelet to Container Runtime
Limits summary
Limits by level
Container Primitives
CFS Quota? Period?
CRI Runtime vs OCI Runtime
Pod Autoscaling
Vertical Pod Autoscaler (VPA)
Conclusion