Learn at your own level
Kubernetes can make sense—one layer at a time.
Start with a familiar port story, then gradually reveal objects, architecture, and production request mechanics. No Kubernetes background required.
- 5guided flows
- 13components
- 0external dependencies
⎈Desired stateAPI-driven control loop
APINode
NetworkData
Your learning path
How deep should we go?
Pick the language that feels comfortable. You can move between levels at any time; your choice is remembered on this device.
Foundation · 1 of 4For curious beginners and non-technical teammates
A familiar story
Imagine a busy shipping port
Your application is cargo that must stay available. Kubernetes is the port authority coordinating where it runs and replacing it when something fails.
One thing to rememberYou describe the result you want; Kubernetes keeps working to maintain it.
Plain words → Kubernetes words
Terminology bridge
Shared foundation
The 60-second mental model
You declare what should exist. Kubernetes continuously compares that desired state with reality and acts to close the gap.
01Declare
Submit an object such as a Deployment through the Kubernetes API.
→
02Store
The API server validates the request and persists cluster state in etcd.
→
03Reconcile
Controllers and the scheduler decide which actions are needed.
→
04Run & observe
Kubelets run Pods and report status. The loop never stops.
Architecture explorer
Who does what?
Select a component to see its responsibility, interactions, and a plain-language analogy.
Cluster edge & add-ons
Control plane · decides
Worker node · runs
Core component Add-on or implementation-dependent
Guided flows
Follow the signal
Choose a real cluster event, then step through the systems involved.
Step 1 of 6Control plane
You + kubectl
Submit desired state
Object relationships
From Deployment to traffic
These resources form a common stateless web application stack. Select an object for its job and lifecycle.
→
→
←
←
Workload controllerDeployment
⚿Configuration
ConfigMap stores non-confidential settings. Secret stores sensitive data, but requires appropriate encryption and access controls.
▤Persistent data
A PersistentVolumeClaim asks for storage. A compatible volume is bound and mounted into the Pod.
⌁Networking
Services select Pods by labels. DNS provides stable names; NetworkPolicy can restrict allowed connections when supported by the network plugin.
↗Scaling
HorizontalPodAutoscaler adjusts replica count from observed metrics. Resource requests matter because utilization is calculated against them.
Read the cluster
A small kubectl toolbox
Avoid confusion
Common misconceptions
“A Service runs my application.”
No. Pods run containers. A Service supplies a stable virtual endpoint and selects eligible Pods.
“Ingress works after I create the YAML.”
An Ingress resource needs an Ingress controller. For new designs, also evaluate the Kubernetes Gateway API.
“A Secret is automatically encrypted.”
Secret data is base64-encoded, not inherently encrypted. Configure encryption at rest and least-privilege RBAC.
“Kubernetes adds persistence to a container.”
Container filesystems are ephemeral. Stateful data needs an appropriate volume, claim, backup, and lifecycle plan.
Continue learning
Authoritative references
The explanations in this app are based on official Kubernetes documentation. Open these for precise behavior and current API guidance.
ArchitectureKubernetes componentsControl plane, node components, and add-ons ↗
WorkloadsPods and controllersDeployments, Jobs, StatefulSets, and more ↗
NetworkingServices and trafficCluster networking, DNS, Gateway, and Ingress ↗
StorageVolumes and claimsPersistent storage concepts and lifecycle ↗
ConfigurationConfigure workloadsConfigMaps, Secrets, probes, and resources ↗
ScalingHorizontal autoscalingHow HPA observes and changes replicas ↗
Hands-onKubernetes BasicsOfficial interactive learning path ↗
CLIkubectl quick referenceEveryday commands and patterns ↗