Radar is an open-source tool that gives AI coding agents context about your Kubernetes cluster. It exposes ranked issues, a change timeline, logs, and service owners over MCP so agents like Claude Code, Codex, or Cursor can investigate failing workloads read-only and report root causes with supporting evidence. Radar Cloud additionally investigates alerts and posts findings to Slack.
Radar
You don't need an AI SRE. Give your agent Radar.

Radar Introduction
Key Features
- Serves Kubernetes cluster context to AI agents over MCP
- Ranked issues, change timeline, logs, and owners in a single view
- One-click Investigate on failing workloads for read-only root cause analysis with evidence
- Open source core with a Radar Cloud option that investigates alerts and posts findings to Slack
Use Cases
- Debugging Kubernetes incidents with AI coding agents
- Investigating failing workloads and identifying root causes
- Routing alert findings to engineering teams via Slack
- Giving agents cluster-specific context they would otherwise lack
Why Startups Use It
Startups running Kubernetes can use Radar to let their existing AI coding agents diagnose cluster issues with real context, turning alerts into evidence-backed root cause analysis without building that tooling in-house. The open-source core makes it easy to evaluate alongside existing agent workflows.
Alternative Options
K8sGPT, HolmesGPT, Klarna AI
Frequently Asked Questions
What is Radar?
Radar is an open-source tool that gives AI agents context about your Kubernetes cluster, including ranked issues, a change timeline, logs, and owners, delivered over MCP.
Which AI agents does Radar work with?
The listing names Claude Code, Codex, and Cursor as agents that can use Radar to investigate failing workloads.
Is Radar free?
The core product is described as open source. No pricing details are provided in the listing, so users should check the official site for current information.
More About Radar
Add our badge to your website to showcase product credibility and listing status.