Freelens: Taking back control of your Kubernetes clusters with a truly open-source desktop client
Managing Kubernetes clusters entirely from the command line is a rite of passage. We have all typed kubectl get pods
Kgateway is a feature-rich, fast, and flexible Kubernetes-native ingress controller and next-generation API gateway. Built on top of the robust Envoy proxy and the Kubernetes Gateway API, Kgateway acts as a reverse proxy, providing a crucial security barrier between your clients and the microservices that constitute your
Kgateway is a feature-rich, fast, and flexible Kubernetes-native ingress controller and next-generation API gateway. Built on top of the robust Envoy proxy and the Kubernetes Gateway API, Kgateway acts as a reverse proxy, providing a crucial security barrier between your clients and the microservices that constitute your application.
Kgateway isn't just another API gateway; it's fully compliant with the Kubernetes Gateway API and extends its functionality with custom Gateway APIs like RoutePolicies, ListenerPolicies, and Backends. These resources allow for centralized configuration of advanced traffic management, security, and resiliency rules for an HTTPRoute or Gateway listener. In simpler terms, Kgateway gives you more control and flexibility over how traffic flows through your Kubernetes cluster.

The kgateway project offers a range of extensions on top of the Kubernetes Gateway API, enabling advanced routing, security, and resiliency capabilities. Some of these extensions include:
When you create a Kubernetes Gateway resource, Kgateway automatically spins up, bootstraps, and manages gateway proxy deployments. This magic is achieved through a combination of Kgateway and Kubernetes resources, including GatewayClass, GatewayParameters, and a gateway proxy template that includes the Envoy configuration for each proxy.
For a deeper dive into the default setup and how these resources interact, check out the Default gateway proxy setup.
Example of a GatewayClass:
apiVersion: gateway.networking.k8s.io/v1
kind: GatewayClass
metadata:
name: kgateway
spec:
controllerName: kgateway.dev/kgateway
description: KGateway Controller
parametersRef:
group: gateway.kgateway.dev
kind: GatewayParameters
name: kgateway
namespace: kgateway-systemKgateway's flexibility allows you to deploy it in a way that best suits your environment. Here are some recommended deployment patterns:
In this setup, a single Kgateway proxy serves as the ingress API gateway for all workloads in a Kubernetes cluster. It's centrally managed by the Kgateway control plane and configured to match and forward traffic based on your defined rules. This is a great starting point for smaller environments where all workloads run in a single cluster.
For larger environments or those with both high and low traffic services, a sharded gateway can help isolate services and protect against noisy neighbors. Multiple gateway proxies split the traffic for different services, providing better load balancing and isolation.
This pattern combines a central ingress gateway proxy with a second layer of sharded gateway proxies. The central gateway applies common traffic management, resiliency, and security rules, while the second layer handles traffic for specific apps, teams, or namespaces. This is ideal if you need a central IP address and DNS name for the gateway that serves all your traffic.
Kgateway can also be paired with other proxy types, such as HAProxy or AWS NLB/ALB, as your central ingress endpoint.
Kgateway can be deployed as an ingress, egress, or waypoint proxy gateway for workloads in an Istio ambient mesh. This allows you to leverage Kgateway's features within your service mesh environment.
For more details, refer to the guides for using Kgateway as an ingress or waypoint proxy for your ambient mesh.
Kgateway is a powerful and versatile tool for managing traffic in your Kubernetes environment. Whether you're running a small cluster or a large, complex infrastructure, Kgateway provides the features and flexibility you need to ensure secure, reliable, and efficient communication between your services. So, go ahead and give Kgateway a try – your microservices will thank you!
Managing Kubernetes clusters entirely from the command line is a rite of passage. We have all typed kubectl get pods
Ya sea que necesites gestionar un millón de H100s o solo quieras asegurarte de que tu agente de IA no tire un rm -rf / en producción, las actualizaciones del Next '26 de GKE sugieren que el futuro de la IA es, inevitablemente, un archivo YAML.