Repolode
Submit
skyhook-io

skyhook-io/radar

A free, open-source Kubernetes UI that runs as a single binary and includes a built-in MCP server so AI agents can read cluster state.

Open-source alternative to:LensOctant

View on GitHubWebsite
Stars
4,062
Forks
268
Open issues
120
Last commit
21 hours ago

What is it?

Radar is a Kubernetes web UI for finding what is broken in a cluster and why. It surfaces a ranked list of current issues such as crash loops, image pull failures, stuck rollouts and dangling references, and lets you trace a resource through its topology, a timeline of events and changes with diffs, Helm releases, GitOps status and live service traffic.

It runs locally as a kubectl plugin or desktop app using your existing kubeconfig, with no agents or CRDs installed in the cluster. It can also be deployed in-cluster with Helm for shared team access, with OIDC or auth-proxy login and per-user Kubernetes RBAC.

Radar includes an MCP server with read-only and full endpoints, so an AI agent can query the same relationships, health data and events the UI shows. It can also run AI investigations through a locally installed agent CLI such as Claude Code or Codex.

Key features

  • 01Ranked issues view for crash loops, image pull errors, stuck rollouts and failing Argo CD or Flux conditions
  • 02Interactive topology graph and timeline of events and resource changes with diffs
  • 03Helm release management, including viewing values, diffing revisions, upgrading and rolling back
  • 04GitOps fleet view for Argo CD and Flux with sync and drift diagnosis
  • 05Traffic view backed by Hubble, Istio, Caretta or Beyla
  • 06MCP server with read-only and full endpoints for AI agents, with secret values withheld from reads

Use cases

  • Triaging a failing deployment by following its topology, events and recent changes in one place
  • Letting an AI agent diagnose cluster problems using live cluster data
  • Giving a team shared, RBAC-scoped access to cluster state through an in-cluster install
  • Auditing workloads for missing probes, resource limits and other best-practice gaps

Quick start

  1. 1

    Install Radar with the official install script from the README

  2. 2

    Install via Homebrew

    brew install kubectl-radar
  3. 3

    Run Radar against your cluster

    kubectl radar

Things to know

  • Most advanced features like AI investigations need an agent CLI installed and signed in on your machine.
  • AI investigations do not run in in-cluster installs and need authentication turned off.
  • Some AI investigation setups let the agent use its own tools and MCP servers, which Radar cannot constrain.

In collections

Open-source AI agent frameworks & appsMCP serversOpenTelemetry & observability platformsMicroservice frameworks & service infrastructure

Topics

#kubernetes#kubernetes-ui#mcp#ai-agents#gitops#helm#observability#troubleshooting#argocd#cloud-native

FAQ

What is radar?
A free, open-source Kubernetes UI that runs as a single binary and includes a built-in MCP server so AI agents can read cluster state. Radar is a Kubernetes web UI for finding what is broken in a cluster and why. It surfaces a ranked list of current issues such as crash loops, image pull failures, stuck rollouts and dangling references, and lets you trace a resource through its topology, a timeline of events and changes with diffs, Helm releases, GitOps status and live service traffic.
Is radar free to use? What license does it have?
radar is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is radar still maintained?
Actively maintained. The last commit to radar was on Nov 6, 2026.
What can I use radar for?
Triaging a failing deployment by following its topology, events and recent changes in one place Letting an AI agent diagnose cluster problems using live cluster data Giving a team shared, RBAC-scoped access to cluster state through an in-cluster install Auditing workloads for missing probes, resource limits and other best-practice gaps
How do I install radar?
Install Radar with the official install script from the README Install via Homebrew: brew install kubectl-radar Run Radar against your cluster: kubectl radar
Where can I get radar?
The source code and installation instructions are on GitHub: https://github.com/skyhook-io/radar

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects