wiamsuri/kube-resource-view repository overview

⁠Kube Resource View

Kube Resource View is an ultra-lightweight, zero-dependency, real-time visualization dashboard designed to give cluster administrators an instant, intuitive look at CPU and Memory allocation and consumption.

Unlike heavy monitoring setups (like Prometheus, Grafana, or OpenTelemetry) that take hours to configure, require persistent storage, and consume substantial cluster resources themselves, Kube Resource View is stateless, instant, and deployable in seconds.

⁠Why Kube Resource View?

As a Kubernetes administrator, answering simple questions about cluster capacity shouldn't require complex query languages or heavy dashboards:

• Which nodes are over-committed on requests, but idle on actual usage?
• Which pods are hitting their memory limits and risking OOM kills?
• How much headroom do we actually have left in the cluster?

Kube Resource View answers these questions at a glance by visually pairing three critical metrics for every node and pod:

1. Actual Consumption (Live metrics from Kubernetes Metrics Server)
2. Resource Requests (Guaranteed capacity)
3. Resource Limits (Max capacity boundaries)

⁠Key Benefits for Cluster Admins

• Zero-Configuration Deployment: No databases to provision, no persistent storage volumes, and no custom resource definitions (CRDs). You deploy a single YAML manifest, and the dashboard is live.
• Instant visual feedback: Dynamic, responsive gauges contrast requests, limits, and actual usage side-by-side, making resource sizing imbalances instantly obvious.
• Real-Time Stream: Powered by Kubernetes watch events and server-sent metrics, pod creations, deletions, and resource usage changes stream directly to your browser as they happen.
• Ultra-Lightweight: The backend is completely stateless and consumes very little memory (~60MB).

⁠One-Command Deployment

You can deploy Kube Resource View directly to your cluster using the pre-configured deployment manifest.

kubectl apply -f k8s/example-deployment.yaml

Once applied, this creates:

1. A ServiceAccount & ClusterRole with read-only access to query nodes, pods, and metrics.
2. A Deployment with 1 replica of the dashboard.
3. A Service exposing the HTTP dashboard.

Note: The cluster must have the standard Kubernetes metrics-server installed for actual CPU & Memory consumption data to populate.

⁠Local Testing (Running via Kubeconfig)

If you prefer to run it locally on your machine (using your current active kubectl context) without deploying it to the cluster:

1. Install dependencies:

   npm install
   

   Copy

2. Start the local server:

   npm run dev
   

   Copy

   Open http://localhost:5173 to see your active cluster immediately.

⁠License

This project is licensed under the MIT License⁠.