Kubernetes · Local · Free

Practice Kubernetes
on a real cluster.

FeatherLab runs a kind cluster on your machine, presents guided labs in your browser, and grades your work against actual Kubernetes state — not the commands you typed.

npm install -g featherlab
View on GitHub npm package

Free and open source · MIT license

The experience

Inside a lab

Every FeatherLab exercise follows the same five steps — from browsing the curriculum to unlocking the next challenge.

Step 1

Open the launcher, pick an exercise.

The launcher shows your progress across all eight levels. Completed exercises stay unlocked; prerequisites gate the next challenge. Resume where you left off.

FeatherLab launcher showing levels 0–7 with progress tracking
Step 2

Read the scenario and requirements.

Each exercise opens with a short scenario, a requirements checklist, and optional hints. For "Your first Pod": create a Pod named web running nginx:1.27 in namespace fl-k8s-0001.

Exercise workspace showing the scenario panel and requirements checklist
Step 3

Work in the integrated terminal.

The terminal is a real shell inside a workstation container joined to your practice cluster. Run kubectl, edit YAML, inspect events — whatever it takes.

Terminal session: kubectl run web --image=nginx:1.27 -n fl-k8s-0001 — pod/web created — kubectl get pods -n fl-k8s-0001 — NAME READY STATUS RESTARTS AGE — web 1/1 Running 0 12s

Step 4

Each check reports what it found.

FeatherLab inspects actual cluster state: does the Pod exist, is the image right, is the phase Running, is the Ready condition True? Each check tells you what it found — not just "wrong".

Grading results showing 4/4 checks passed with individual check feedback
Step 5

Completion unlocks the next exercise.

Progress is stored locally in SQLite and survives restarts. Completing an exercise unlocks what follows. Move on when you're ready.

Exercise completed
Next: Kubernetes Basics

Opens in a fresh environment: this exercise's cluster resources are reset. Your completion stays saved.

What makes it different

Graded on cluster state,
not the commands you typed.

FeatherLab validates Kubernetes objects directly — it reads the Pod's phase, checks the image field, and evaluates every condition. Two different commands that produce the same result both pass. Each check reports exactly what it found, not just "wrong". Partial credit is tracked: a score of 3/4 tells you which condition still needs work.

If an infrastructure failure prevents a check from running, the outcome is incomplete — distinct from a wrong answer and never recorded as a failed attempt.

⌨kubectl
⎈cluster
◈validators
✓ Pod exists in fl-k8s-0001
✓ Container uses nginx:1.27
✓ Pod is Running
✓ Pod Ready condition is True
No account required
Runs entirely on your machine
127.0.0.1 only
Server never binds to a public address
Progress in SQLite
Stored locally in ~/.featherlab
Resume safely
Cluster and workstation survive featherlab stop

What you can learn

174 exercises across 8 levels — foundations to NetworkPolicy.

Each level introduces one concept at a time, then applies it, then breaks it on purpose. Guided labs, independent exercises, troubleshooting challenges, and mastery activities are interleaved with knowledge checks throughout each level.

Level 0

Kubernetes foundations

10 exercises7 knowledge checks
Level 1

Deployment lifecycle

18 exercises6 knowledge checks
Level 2

Configuration, health, recovery

25 exercises6 knowledge checks
Level 3

Application patterns

25 exercises7 knowledge checks
Level 4

Resource management & scheduling

19 exercises8 knowledge checks
Level 5

Services and DNS

27 exercises7 knowledge checks
Level 6

Ingress and Gateway API

26 exercises6 knowledge checks
Level 7

NetworkPolicy

24 exercises7 knowledge checks

In development

Level 8
Storage — PVs, PVCs, StorageClasses
Level 9
ServiceAccounts and RBAC
Level 10
Kustomize, Helm, and CRDs
Levels 11–12
CKAD mastery and troubleshooting
Levels 13–15
Cluster administration and CKA readiness
Levels 16–21
Cloud-native security and CKS readiness

FeatherLab is being developed toward CKA, CKAD, CKS, KCNA, and KCSA coverage. Not affiliated with or endorsed by the CNCF.

Curriculum references

Topic coverage only — no exercise text, manifests, or questions reproduced.

Quick start

Up and running in three commands.

FeatherLab is an npm package. Install it globally, verify prerequisites, and launch.

Before you start

  • Linux — macOS is exploratory but not regularly tested; Windows and WSL2 are not validated
  • Node.js 22.14 or later
  • Docker Engine — docker info should succeed without sudo
  • kind CLI — version 0.33.0 recommended

Internet access is needed on first launch to pull the kind node image and workstation container. Most subsequent use works offline once images are cached.

01
InstallInstall the featherlab CLI globally from npm.
npm install -g featherlab
02
Check prerequisitesVerifies Docker and kind are reachable.
featherlab doctor
03
LaunchCreates the kind cluster on first run, starts the browser workspace.
featherlab

Opens at http://127.0.0.1:4317.

View on GitHub npm package

Common questions

Frequently asked questions

Is FeatherLab free?
Yes. FeatherLab is completely free and MIT-licensed. Install it with npm install -g featherlab at no cost — no cloud account, no subscription, no credit card required.
What Kubernetes topics can I practice?
Current labs cover pods, deployments, ConfigMaps and Secrets, health checks, resource management and scheduling, Services and DNS, Ingress, Gateway API, and NetworkPolicy — over 170 exercises across 8 levels. The roadmap continues with RBAC, storage, Helm, and CKA, CKAD, and CKS exam readiness.
Do I need a cloud account?
No. FeatherLab runs a real kind (Kubernetes in Docker) cluster entirely on your own machine. You need Docker and Node.js — no cloud provider or account required.
Is FeatherLab open source?
Yes. Source code is on GitHub under the MIT license. Contributions are welcome.
What certifications does FeatherLab help prepare for?
The curriculum is being developed toward CKA, CKAD, CKS, KCNA, and KCSA coverage. FeatherLab is not affiliated with or endorsed by the CNCF.