Closing the Observability Gap: Building a Kubeflow Plugin for Headlamp - Alok Dangre, NIT Rourkela

Why watch

This session demonstrates the Kubeflow Headlamp plugin that integrates Kubeflow CRDs (Pipelines, Runs, Experiments, Notebook servers, Katib experiments, TrainJob, Spark applications) directly into the Headlamp Kubernetes dashboard, eliminating context-switching between UIs. Built with TypeScript using the Headlamp Plugin SDK during an LFX Mentorship, the plugin follows a 'complement, not duplicate' principle providing cluster-level observability with one-click deep links to Kubeflow's native interfaces. Ideal for ML engineers and platform teams managing ML workloads on Kubernetes who want unified cluster visibility.