Storage & Scaling

Requests influence scheduling; limits constrain runtime use. Readiness controls traffic, liveness restarts a stuck container, and startup probes protect slow initialization. This manifest makes all four behaviors explicit.

The HorizontalPodAutoscaler changes replica count from observed metrics. This guide installs Metrics Server, deploys a workload with CPU requests, creates an autoscaler, generates load, and watches scale-up and scale-down.

A StatefulSet gives Pods stable identities and per-replica claims. This single-replica PostgreSQL lab demonstrates a headless Service, Secret-backed password, volumeClaimTemplates, readiness, backup, and controlled deletion.

Scheduling controls decide where a Pod may or should run. This lab labels a node, demonstrates required node affinity, protects a dedicated node with a NoSchedule taint, and adds a matching toleration.

PersistentVolumes represent storage capacity, PersistentVolumeClaims request it, and StorageClasses define dynamic provisioning behavior. This lab uses a hostPath PV only to make the binding lifecycle visible on a single-node lab.

The NFS subdir external provisioner dynamically creates a subdirectory for each PVC on an existing NFS export. This guide validates the export, installs the maintained Helm chart, creates a claim, and tests…