Storing data inside your cluster with StatefulSets

It's best to store data within your Kubernetes cluster. This provides a uniform one-stop shop to manage your workloads and all the resources they depend on (excluding third-party external services). Additionally, you get to integrate your storage with your streamlined monitoring, which is very important. We will discuss monitoring in depth in a future chapter. However, running out of disk space is the bane of many system administrators. But there is a problem if you store data on a node and your data store pods get rescheduled to a different node, and the data it expects to be available is not there. The Kubernetes designers realized that the ephemeral pod philosophy doesn't work for storage. You could try to manage it yourself using pod-node affinity and other mechanisms that Kubernetes provides, but it's much...