Chapter 8: Pod Placement Controls

This chapter describes the various ways of controlling Pod placement in Kubernetes, as well as explaining why it may be a good idea to implement these controls in the first place. Pod placement means controlling which node a Pod is scheduled to in Kubernetes. We start with simple controls like node selectors, and then move on to more complex tools like taints and tolerations, and finish with two beta features, node affinity and inter-Pod affinity/anti-affinity.

In past chapters, we've learned how best to run application Pods on Kubernetes – from coordinating and scaling them using deployments, injecting configuration with ConfigMaps and Secrets, to adding storage with persistent volumes.

Throughout all of this, however, we have always relied on the Kubernetes scheduler to put Pods on the optimal node without giving the scheduler much information about the Pods in question. So far, we've added resource limits and requests to our...