The Kubernetes Ecosystem and Distributions

You rarely run vanilla Kubernetes alone. A rich ecosystem of distributions and add-ons sits on top — knowing the map helps you pick a starting point.

Upstream Kubernetes is the raw open-source project. A distribution wraps it with installers, defaults, and support — much like Linux distros wrap the kernel.

Managed Kubernetes — EKS, GKE, AKS — lets the cloud run your control plane, upgrades, and availability. You just manage your workloads.

For laptops, edge, and CI, reach for lightweight distros: Minikube, kind, k3s, or MicroK8s. Perfect for learning and fast local clusters.

For on-prem and enterprise, self-managed distributions like OpenShift, Rancher, and Tanzu add UI, security defaults, and lifecycle tooling on top.

The CNCF (Cloud Native Computing Foundation) hosts Kubernetes plus its companion projects, and certifies distributions so your workloads stay portable.

Kubernetes ships no networking of its own. You add a CNI plugin — Calico, Cilium, or Flannel — to wire up pod networking and policies.

Storage works the same way: the CSI standard lets vendors plug in. Managed clusters usually hook this up to cloud disks for you automatically.

For visibility and traffic control, pair Prometheus and Grafana for metrics with a service mesh like Istio or Linkerd for routing and mTLS.

Helm is the de-facto package manager. A chart bundles all the manifests so you install a complex app with a single command.

Rule of thumb: managed for production without ops burden, lightweight local for learning and CI, self-managed when you need on-prem control.

Match the use case to the right Kubernetes flavour.

You have the Kubernetes map: managed vs local vs self-managed, the CNCF for conformance, CNI/CSI for networking and storage, and Helm for packaging.