Local Kubernetes Dev — Part 3: Tooling overview — who does what

# devops# docker# kubernetes
Local Kubernetes Dev — Part 3: Tooling overview — who does whatGaberial Sofie

k3d, kind, minikube, Helm, Kustomize, Tilt, Skaffold, DevSpace, Telepresence, k9s... it reads like an...

k3d, kind, minikube, Helm, Kustomize, Tilt, Skaffold, DevSpace, Telepresence, k9s... it reads like an incantation, and it's easy to drown in.

If you've ever opened the CNCF landscape and felt you had to master all of it at once — relax. In reality each tool owns its own little shelf, and once you sort them out, the picture gets simple. Part three of the local Kubernetes series is a map of the 2025–2026 ecosystem: who does what, and why you probably only need two tools from this list.

I lay it all out: Docker and kubectl are the foundation (cluster nodes are just Docker containers); local clusters k3d / kind / minikube / Docker Desktop / Rancher Desktop; Helm vs Kustomize (templating package manager vs overlay patches); inner-loop accelerators Tilt / Skaffold / DevSpace; the separate "remote" approach Telepresence / mirrord / Gefyra; and k9s as a visual kubectl. And I honestly explain why, for the end-to-end myapp example (FastAPI + PostgreSQL), the series picked k3d + Tilt: a real Kubernetes API with a fast start plus Live Update that squeezes the loop from minutes to seconds.

Opinionated, but honestly so: "we chose X for these reasons," not "X is the best." https://dorokhovich.com/blog/local-k8s-tooling-overview?utm_source=devto&utm_medium=syndication&utm_campaign=local-k8s-tooling-overview