Documentation
Everything a DevOps engineer needs to install, configure and operate krouter.
krouter is a Kubernetes Gateway API implementation and reverse proxy. It routes HTTP, gRPC, TCP, UDP and TLS traffic, is configured entirely through standard Kubernetes resources (no custom CRDs), and applies every configuration change without dropping connections.
If you are new to the Gateway API itself, start with the upstream API overview: the Gateway, HTTPRoute and GatewayClass concepts carry over to krouter unchanged. General Kubernetes concepts used throughout these pages are covered by the official Kubernetes documentation.
How krouter is put together
- A control plane (single-replica Deployment) watches Gateway API resources, validates them, compiles configuration and publishes status. It also serves the live dashboard.
- A data plane (DaemonSet) serves the actual traffic on every node. Data-plane pods never talk to the Kubernetes API for routing decisions on the hot path.
- For every Gateway, krouter generates a regular Service (NodePort by default) fronting the shared data plane.
Configuration changes are compiled into immutable generations and applied atomically: a broken change is rejected in full and the last valid configuration keeps serving.
Where to go next
- Installation: prerequisites and deployment.
- Configuration: GatewayClass, parameters, listeners and TLS material.
- Routing: route kinds, matching, filters and cross-namespace policies.
- Extensions: per-rule rate limiting and Coraza WAF from plain ConfigMaps.
- Authentication: per-rule OIDC, SAML, LDAP and JWT authentication from plain Secrets.
- Observability: dashboard, metrics and logs.
- Conformance & testing: how krouter is verified, including the full test report.
- Tutorials: hands-on walkthroughs from a first route to full mutual TLS.