✦ For everyone, free.

Practical knowledge for real and everyday life

Home

1.3 Kubernetes Control Plane Definition

The Kubernetes Control Plane manages cluster operations, orchestrating containers, scheduling workloads, and enabling service discovery.

Kubernetes Control Plane Definition is the precise characterization of the control plane as the decision-making subsystem of a Kubernetes cluster, distinguished from worker nodes by the fact that it never executes application workloads itself, existing solely to accept declared state, persist it durably, and drive the cluster toward matching that state.


Defining the Control Plane by Its Function

What It Does, Not Where It Runs

The control plane is best defined by its function rather than its physical location: it is whatever set of processes accepts API requests, stores the authoritative cluster state, schedules workloads onto nodes, and runs the reconciliation loops that keep the cluster's actual state aligned with its desired state.

Distinct from Worker Capacity

Unlike worker nodes, which contribute compute capacity for running Pods, the control plane's role is purely coordinative; even in small clusters where control plane processes may run on the same physical machines as workloads, the roles remain conceptually and functionally distinct.

Control Plane = coordination execution

The Minimal Set of Responsibilities

Accepting and Storing Desired State

At a minimum, the control plane must expose an API through which desired state can be declared, and it must persist that state durably so it survives restarts of any individual control plane process.

Reconciling State Continuously

Beyond storage, the control plane is defined by its continuous reconciliation behavior: watching the difference between desired and actual state, and issuing the necessary actions, such as scheduling decisions or Pod recreation, to close that gap.


What Is Not Part of the Control Plane

Application Workloads

Application containers, regardless of how critical they are to a business, do not run as part of the control plane; they run on worker nodes and are merely the subject of the control plane's reconciliation decisions.

The Container Runtime

The container runtime, responsible for actually executing containers, is a worker-node-level concern and is not considered part of the control plane, since the control plane never invokes it directly.


Why a Precise Definition Matters

Clarifying Where Responsibility Lies

A precise definition of the control plane clarifies operational responsibility: issues with scheduling, API availability, or cluster-wide state consistency are control plane concerns, while issues with an individual container crashing are typically node- or application-level concerns instead.

Informing High Availability Design

Because the control plane is defined by its coordinative function rather than any single implementation, high availability design for it focuses on replicating the components that fulfill that function, such as the API server and the cluster's data store, rather than replicating any particular machine.


Control Plane Boundary Diagram

Control Plane API, store, scheduler, controllers Worker Nodes kubelet, runtime, running Pods