Advanced Pod Settings

Previous sections cover Pod basics explaining what a Pod is and how to declaratively describe one.

It has been described how a Pod is the smallest schedulable unit in Kubernetes, may contain one or several InitContainers and Containers, that they may share volumes and share a loopback network device with a common port range.

In this section, more advanced Pod settings are discussed such as Pod scheduling and health monitoring.

When provisioning one or even multiple distributed systems to a single Kubernetes cluster, many microservices will co-exist within the cluster potentionally competing for Kubernetes Node resources. It is therefore valuable to understand how the Kubernetes scheduler places Pods among Kubernetes Nodes and what happens if resources become scarce and Pods fail.

Kubernetes has built-in health-monitoring capabilities. In previous sections, the self-healing of failed Pods has already been studied. In this section, we'll have a closer look at how the health and readiness of a container are determined. More than that, we'll see that health and readiness probes can be customized to context specific needs.