K3s

Go up to Terms and Definitions

K3s is a lightweight Kubernetes distribution designed for resource-constrained, edge, remote, development, and production environments.

K3s packages Kubernetes control-plane operation into a simplified distribution with reduced external dependencies. K3s uses a server-and-agent architecture: a server node runs control-plane and datastore components managed by K3s, while an agent node runs without datastore or control-plane components. Both server and agent nodes run the kubelet, container runtime, and container networking interface.

In the Financial Systems Archetype document set, K3s belongs to implementation, deployment, and runtime-infrastructure concerns. K3s does not define logical Nodes, logical Communication Endpoints, logical information structures, logical Runtime Planes, or logical interaction patterns.

An implementation profile maps selected logical Nodes, communication responsibilities, configuration concerns, observability concerns, and deployment concerns to K3s resources when K3s serves as the selected Kubernetes distribution.

lightweight Kubernetes distribution used to deploy, operate, and manage containerised workloads in resource-constrained, edge, remote, development, or production environments

Generalised from the official K3s documentation and K3s architecture description, and its use as a deployment concern in Part 4: Phase 0 Implementation Profile / PSM and Part 5: Phase 0 Developer Handbook. K3s describes itself as a certified Kubernetes distribution for production workloads in unattended, resource-constrained, remote, IoT, and edge environments. Its architecture defines server nodes and agent nodes, with both running kubelet, container runtime, and container networking interface components.

K3s is not Kubernetes itself, a logical Node, logical Communication Endpoint, Runtime Plane, DDS Domain, DDS Topic, message bus, data structure, governance model, evidence model, or architecture layer.

Part 5 uses K3s as an optional deployment target for running containerised Phase 0 nodes after the logical Nodes and logical Communication Endpoints have already been defined in platform-independent form.