Skip to main content

Practice exam · CNCF

Kubernetes and Cloud Native Associate (KCNA) practice exam

50
questions
90
minutes
75%
to pass the practice exam
60
questions in the bank

Short answer

The Kubernetes and Cloud Native Associate (KCNA) practice exam has 50 questions in 90 minutes, with a 75% passing mark, as configured here. The fastest way to know whether you are ready is to sit a full exam on the clock and read your score by domain, not only the total.

Domains covered

  • Cloud Native Architecture
  • Cloud Native Application Delivery
  • Cloud Native Observability
  • Container Orchestration
  • Kubernetes Fundamentals

Sample questions, with explanations

Three questions from the bank, open to read.

Question 1 · Kubernetes Fundamentals

A team needs to run 5 identical replicas of a stateless web app and wants Kubernetes to automatically recreate Pods if they crash and to support rolling updates. Which resource is the most appropriate choice to declare and manage this desired state?

  1. A standalone Pod, which Kubernetes scales automatically when it crashes
  2. A Deployment, which manages a ReplicaSet to maintain the desired number of Pod replicas ✓ (correct)
  3. A Namespace, which controls how many replicas of an app may run
  4. A Service, which keeps the desired count of replicas running and restarts failed Pods
Why this is correct

A Deployment, which manages a ReplicaSet to maintain the desired number of Pod replicas, is correct: Deployments are the standard way to declaratively manage stateless replicated apps, providing self-healing (the underlying ReplicaSet recreates failed Pods) and rolling updates. 'A Service, which keeps the desired count of replicas running and restarts failed Pods' is wrong because a Service only provides stable network access and load balancing to Pods, it has no role in replica count or restarts. 'A standalone Pod, which Kubernetes scales automatically when it crashes' is wrong because a bare Pod is not self-healing or scalable, if its node fails the Pod is gone with no controller to recreate it. 'A Namespace, which controls how many replicas of an app may run' is wrong because a Namespace is a logical partition for organizing resources, not a replica controller (quota limits exist but are separate ResourceQuota objects, not the replica manager).

Question 2 · Cloud Native Architecture

In a standard Kubernetes control plane, which CNCF project serves as the distributed key-value store that holds the cluster's entire state and configuration data?

  1. Prometheus, which is responsible for storing the cluster's state and configuration data
  2. etcd, a consistent and highly available key-value store used as Kubernetes' backing store ✓ (correct)
  3. containerd, which stores cluster state in addition to running the containers
  4. Envoy, which acts as the persistent database for all Kubernetes object definitions
Why this is correct

etcd is correct because Kubernetes uses it as the single source of truth for all cluster state, storing every object's desired and current configuration in a consistent key-value store. Prometheus is wrong because it is a monitoring and time-series metrics system; it scrapes and stores metrics, not cluster object state. Envoy is wrong because it is an L7 proxy used for service networking and as a data-plane proxy in service meshes, not a database. containerd is wrong because it is a container runtime that pulls images and runs containers; it does not store cluster state.

Question 3 · Container Orchestration

Kubernetes is designed so that it can work with different container runtimes (containerd, CRI-O, and others) without changing its own core code. What makes this pluggability possible?

  1. Each runtime being required to be a fork of Docker so the commands stay identical
  2. The Container Runtime Interface (CRI), a standard API the kubelet uses to talk to any compliant runtime ✓ (correct)
  3. Kubernetes embedding a single fixed runtime that all clusters must use
  4. The Container Network Interface (CNI), which standardizes how containers are scheduled
Why this is correct

The Container Runtime Interface (CRI), a standard API the kubelet uses to talk to any compliant runtime is correct because CRI defines a common contract so the kubelet can drive any runtime that implements it, making runtimes interchangeable. The Container Network Interface (CNI) is wrong because CNI standardizes pod networking, not runtime communication, and it has nothing to do with scheduling. Requiring each runtime to be a fork of Docker is a misconception: runtimes like containerd and CRI-O are independent projects, and CRI is what lets them plug in, not shared lineage. Kubernetes embedding a single fixed runtime is wrong because that would be the opposite of pluggability; CRI exists precisely so no single runtime is hardcoded.

Take the full practice exam

  • 50 questions drawn from the bank, with a 90-minute timer
  • Score by domain and a review of every question
  • An AI tutor to talk through each question you miss
  • What you miss feeds your study plan
Start the practice examStudy the path first

Independent, unofficial practice exam. CNCF and the exam name are trademarks of their owners; UpSkillScore is not affiliated with them. Format and passing mark are the exam's configuration in UpSkillScore, not an official statement.

Frequently asked questions

How many questions does the Kubernetes and Cloud Native Associate (KCNA) practice exam have?

It has 50 questions and a 90-minute limit. The passing mark of the practice exam is 75%.

Is this the official Kubernetes and Cloud Native Associate (KCNA) exam?

No. It is an independent practice exam built to match the structure of the real one. It does not replace the official exam.

Do I need an account to take it?

You can read the structure and sample questions without an account. Taking the full timed exam, scoring by domain and the AI tutor need one.