OnCallReadyStart learning

Security and secrets Lesson 23.1 · Kubernetes Security (CKS) · 2 min read

The CKS map: what a cluster attacker goes after

In plain words

A castle is not safe because its gate is strong. It has walls, guards at the door, locked rooms inside and watchmen on the towers, and each of them assumes the one before might fail. A Kubernetes cluster is the same: the workloads, the API server, etcd, the nodes, the images and the running containers each need their own defence, because one strong layer alone always has a gap somewhere.

Why a second security chapter

The problem. A pod that runs as non-root with a read-only root filesystem (chapter 19) is a good start, but an attacker who gets credentials, an unsigned image or a shell in a container does not care about one pod. They read Secrets straight out of etcd, call the API with a stolen token, or run a miner in a container nobody watches.

The exam. The Certified Kubernetes Security Specialist (CKS) is a 2-hour hands-on exam on Kubernetes v1.35 that needs a passed CKA. Its six domains: Cluster Setup (15%), Cluster Hardening (15%), System Hardening (10%), Minimize Microservice Vulnerabilities (20%), Supply Chain Security (20%), and Monitoring, Logging and Runtime Security (20%) (the Linux Foundation's exam page, checked 21 Oct 2026). Check it again before you book: the curriculum moves with each Kubernetes release.

LayerQuestionThis course
WorkloadCan a pod do more than it needs?chapter 19: PSA, securityContext, RBAC
API serverWho did what, and is it allowed?audit logging (next lesson), RBAC
etcdAre Secrets readable at rest?encryption at rest
Nodes and componentsAre the control plane flags safe?kube-bench (CIS)
ImagesIs this the image we built, without known holes?trivy, cosign, Kyverno
RuntimeIs something doing what it never should?Falco

In an interview: defence in depth means each layer assumes the one before it failed: signed images still get scanned, admitted pods still get watched at runtime.

On the exam: most tasks are edits to a static pod manifest on the control plane, a policy object, or a node file, followed by proving it works. That is exactly how these labs run.

Why it helps

The CKS exam and real security reviews are organised by those layers. Knowing which control protects which layer tells you where a gap is when you look at a cluster, and it gives you a clear structure when an interviewer asks how you would secure a cluster: you walk through the layers instead of listing tools at random. It also shows what earlier chapters already covered and what this one adds.

On the blog

FAQ

What are the CKS domains?

Cluster setup, cluster hardening, system hardening, minimize microservice vulnerabilities, supply chain security, and monitoring, logging and runtime security. The last three carry the most weight, about a fifth of the exam each, so the supply chain and runtime lessons deserve extra practice.

What does defence in depth mean here?

Every layer assumes the one before it failed: signed images still get scanned, admitted pods still get watched at runtime, and an RBAC mistake is still visible in the audit log. No single control is trusted to stop everything, so a gap in one is caught by another.

Where does chapter 19 stop?

At single workloads: Pod Security Standards, securityContext, RBAC and Secrets. This chapter secures the cluster around those workloads - the control plane, etcd, the nodes, the images and the runtime - and links back to chapter 19 instead of repeating it.

Is the CKS exam multiple choice?

No. It is a two-hour hands-on exam on real clusters where you type the commands and edit the files yourself, and it requires a passed CKA. That is why every lesson here ends in a lab where you prove the control works.

Where is the official curriculum?

In the CNCF curriculum repository on GitHub, github.com/cncf/curriculum. Check it before you book, because the domains and the Kubernetes version are updated as Kubernetes releases move on, and the lessons here follow the domain list.

In an interview Mid

What are the layers of Kubernetes security?

Workloads (Pod Security, securityContext, RBAC), the API server (authentication, RBAC, audit logging), etcd (encryption at rest), nodes and components (CIS benchmark checks, verified binaries, AppArmor, seccomp, the metadata endpoint), images (scanning, signing, admission control) and runtime (detection with Falco). It is defence in depth: each layer assumes the one before it failed, so one gap does not hand over the cluster.

Also asked: What does defence in depth mean for a cluster? · Which CKS domain carries the most weight? · How is the CKS exam run?

Practise this lesson in the terminal Free, in your browser - a real Ubuntu terminal to try it in, with missions that check your work.