What is Minikube, and why shouldn't you use it in production?
Quick Answer
Minikube runs a single-node Kubernetes cluster locally (in a VM or container) for learning and development. It's not for production because it's single-node — no real high availability, no multi-node scheduling, and limited scale.
Detailed Answer
Minikube spins up a full but single-node Kubernetes cluster on your laptop, using a driver (Docker, HyperKit, KVM, etc.), so you get a real API server, scheduler, and kubelet to develop against. It's great for learning, testing manifests, and CI. It is explicitly not a production tool: a single node means no fault tolerance (the control plane and your workloads share one machine), you can't test real multi-node scheduling, networking, or node failure, and it doesn't scale. For production you use a managed cluster (EKS/GKE/AKS) or a multi-node self-managed cluster (kubeadm, kOps). For local multi-node testing, kind (Kubernetes-in-Docker) is a common alternative. Minikube also bundles handy addons (dashboard, ingress, metrics-server) you enable with one command.
Code Example
minikube start --driver=docker --nodes=1 minikube addons enable ingress minikube addons enable metrics-server kubectl get nodes # a single 'minikube' node
Interview Tip
Say clearly 'single-node = no HA = not production,' and mention kind as the alternative for multi-node local testing. That shows you know the local-dev landscape, not just one tool.