Core concepts
Cluster architecture
Explore the control plane, worker nodes, kubelet, scheduler, and the declarative reconciliation loop.
What you'll learn
- Map control-plane components
- Describe the reconciliation loop
- Trace a workload from API to node
Key idea
The control plane decides; worker nodes execute.
The API server is the front door connecting kubectl, controllers, schedulers, and node agents.
Interactive command lab
Change the syntax. See what Kubernetes does.
Original syntax
kubectl cluster-info [flags]$ kubectl cluster-infoExecution path
Read from top to bottom. Every arrow explains the action and destination.
Outside the cluster · your computer
kubectl
Builds the API request from the command you entered
SEND HTTPS API REQUEST
kubectl sends the desired operation to the Kubernetes API server
KUBERNETES CLUSTER
The control plane manages one or more worker nodes
CONTROL PLANE
Makes cluster-wide decisions; it does not run your application container
API server
Authenticates, validates, and stores the desired state
WATCH FOR UNSCHEDULED PODS
The scheduler watches the API server for Pods without a node
Scheduler
Chooses the best worker node based on resources and constraints
RECONCILE DESIRED STATE
Controllers compare desired state with actual state through the API server
Controllers
Create, replace, or repair managed Kubernetes resources
BIND POD TO WORKER NODE
The scheduling decision is stored; the selected node's kubelet sees the PodSpec
WORKER NODE
A machine that runs application workloads
kubelet
The node agent reads the PodSpec and asks the container runtime to start it
Ready for the guided tour
The tour opens as a fixed panel, so the diagram remains fully visible.
Quick knowledge check
Two-question quiz
Choose one answer for each question
Feedback appears after you check your answers.