Essential Kubernetes Interview Questions and Answers Explained

Kubernetes is a leading container orchestration platform used by many global companies. This article compiles the most important interview questions about Kubernetes, organized into four sections: basic questions, architecture‑focused questions, scenario‑based questions, and multiple‑choice questions.

1. Basic Kubernetes Interview Questions

Q1: Difference between Kubernetes and Docker Swarm – Kubernetes offers a more extensive API, higher scalability, and a richer ecosystem compared to Docker Swarm.

Q2: What is Kubernetes? – An open‑source container management system that handles deployment, scaling, and load‑balancing of containers across clusters.

Q3: Relationship between Kubernetes and Docker – Docker builds and runs containers; Kubernetes orchestrates multiple containers, enabling communication and scheduling across hosts.

Q4: Deploying on a host vs. a container – Host deployment shares the OS kernel and libraries among applications, while containers isolate each application with its own libraries and binaries, preventing interference.

Q5: What is container orchestration? – Coordinating multiple containers (e.g., micro‑services) so they work together, similar to instruments in an orchestra.

Q6: Why is orchestration needed? – Without it, containers cannot communicate reliably; orchestration provides networking, scheduling, and health‑checking.

Q7: Kubernetes features – Includes self‑healing, horizontal scaling, service discovery, and declarative configuration.

Q8: How Kubernetes simplifies container deployment – Provides cloud‑agnostic load‑balancing, scaling, and monitoring across multiple hosts.

Q9: Understanding a Kubernetes cluster – The cluster stores desired state in etcd; the API server receives deployment manifests and schedules pods accordingly.

Q10: Google Container Engine (GKE) – A managed Kubernetes service on Google Cloud.

Q11: Heapster – A cluster‑wide metrics aggregator running as a pod.

Q12: Minikube – A tool to run a single‑node Kubernetes cluster locally.

Q13: Kubectl – The command‑line client for interacting with the Kubernetes API.

Q14: Kubelet – An agent on each node that ensures containers described in a PodSpec are running.

Q15: Node components – Each node runs a kubelet and kube‑proxy, providing networking and pod management.

2. Architecture‑Based Questions

Components of Kubernetes architecture – Master node (kube‑apiserver, controller‑manager, scheduler) and worker nodes (kubelet, kube‑proxy).

Kube‑proxy – Handles TCP/UDP forwarding for services on each node.

Master node responsibilities – Schedules pods, maintains cluster state, and exposes the API.

kube‑apiserver and kube‑scheduler – API server is the front‑end for all cluster operations; scheduler assigns pods to suitable nodes based on resources.

Controller manager – Runs core controllers (e.g., replication, endpoint) that manage the desired state.

etcd – Distributed key‑value store written in Go, holding the cluster configuration and state.

Kubernetes service types – ClusterIP, NodePort, LoadBalancer, and ExternalName.

Load balancer – Internal load balancers balance traffic within the cluster; external load balancers expose services to the internet.

Ingress – API object that defines external access rules (HTTP/HTTPS) for services.

Ingress traffic flow example – Detailed packet path through flannel, bridge, and node networking.

Cloud controller manager – Abstracts cloud‑specific functionality (e.g., storage, networking) from the core Kubernetes code.

Container resource monitoring – Tools like Heapster, Prometheus monitor CPU, memory, and network usage at pod, node, and cluster levels.

ReplicaSet vs. ReplicationController – Both ensure a set number of pod replicas; they differ in selector types (set‑based vs. equality‑based).

Headless Service – Service without a cluster IP, allowing direct pod access.

Kubernetes security best practices – Include RBAC, network policies, secrets management, and regular patching (illustrated with an image).

Cluster federation – Manages multiple clusters as a single entity for multi‑cloud or multi‑region deployments.

3. Scenario‑Based Questions

Various real‑world scenarios explore how companies can transition from monolithic architectures to microservices, manage distributed workloads, achieve scalability, and operate across multiple clouds using Kubernetes. Each scenario provides a concise solution emphasizing containerization, CI/CD pipelines, and Kubernetes features such as scheduling, resource allocation, and multi‑cluster federation.

4. Multiple‑Choice Interview Questions

Minions in a Kubernetes cluster are the worker nodes.

Cluster data is stored in etcd.

Both ReplicaSet and Deployment are Kubernetes controllers.

Pods, Services, and Volumes are all core Kubernetes objects.

The network proxy runs on all nodes.

Node controller responsibilities include CIDR allocation, node list maintenance, and health monitoring.

Replication Controller ensures desired pod count, updates pods, and recreates failed pods.

Defining a service without a selector uses an external name.

Kubernetes 1.8 introduced Taints and Tolerations.

The program checking a container's IP address is TCPSocketAction.

Author: fiisio (translation source: https://zhuanlan.zhihu.com/p/74560934)