Comparison

Radar vs Headlamp

Both Apache 2.0. Different architectures (single Go binary vs Electron + plugins), different scope. Here's where each one wins.

TL;DR
  • Same license. Both Apache 2.0. Headlamp wins on vendor-neutral governance (kubernetes-sigs, CNCF Sandbox); Radar OSS has a public no-relicense pledge.
  • Different architecture. Headlamp = core dashboard + plugin SDK for extensions. Radar = single Go binary; ecosystem features in core, not bolted on as plugins.
  • Different surface. Headlamp covers resource browsing well. Radar adds persistent event timeline, live workload-to-workload traffic with mTLS, cross-cluster topology, and ArgoCD + Flux + OpenCost + Trivy in core (vs separate Headlamp plugins).
  • Pick Radar if you want the operations stack (timeline, traffic, fleet, GitOps, cost) integrated and ready out of the box. Pick Headlamp if vendor-neutral CNCF governance or building on its plugin SDK is the deciding factor.
Feature by feature

The full comparison.

Unmarked Radar checks ship in OSS (Apache 2.0) and Cloud. 'Cloud' tags are hosted-only.

Feature
Radar
OSS (Apache 2.0) + Cloud
Headlamp
kubernetes-sigs · CNCF Sandbox
Project & licensing
License
Apache 2.0
Apache 2.0
Governance
Headlamp is a kubernetes-sigs sub-project, CNCF Sandbox. Radar is built and maintained by Skyhook (commercial OSS) under a public no-relicense pledge.
Skyhook (commercial OSS)
kubernetes-sigs / CNCF Sandbox
Pricing
OSS free, Cloud per cluster
Free
Architecture
Extension model
Headlamp's plugin SDK is its primary surface for adding features. Radar integrates ecosystem features directly into core.
Plugin-less by design
Plugin SDK (TypeScript/React)
Binary
Single Go binary, embedded UI
Electron desktop app + Helm chart
Hosted SaaS option
Cloud
Multi-cluster
Clusters in one view
OSS: multiple clusters per install. Cloud: one fleet topology across all of them. Headlamp added side-by-side compare in 2025 but defaults to one cluster at a time.
Cloud (fleet) · OSS (per install)
Dropdown + side-by-side compare
Cross-cluster resource search
Cloud
Per cluster
Drift detection (staging vs prod)
Cloud
History & incidents
Event timeline retention
Kubernetes API holds events for ~1 hour by default. Headlamp shows what's there. Radar persists past that.
Configurable, persisted
None (live state only)
Post-mortem replay
Audit log retention
Cloud · 7 days to unlimited
GitOps correlation
ArgoCD + Flux, in core
Flux plugin only
Auth
OIDC SSO
DIY (Keycloak/Dex/OpenUnison docs)
SAML SSO
Cloud
SCIM provisioning
Enterprise
Inherits Kubernetes RBAC
Feature surface
Resource browsing + RBAC viewer
Logs, exec, terminal
Live workload-to-workload traffic + mTLS
Topology graph
Live workload graph
Resource map
Image filesystem viewer
Cost insights (OpenCost)
Both require OpenCost installed in the cluster. Radar auto-detects it and renders cost against workloads. Headlamp's flow is install + configure the OpenCost plugin separately.
Auto-detected
OpenCost plugin
Cluster security audit (Trivy / Kyverno)
MCP for AI clients
Radar exposes the cluster as an MCP endpoint your existing Claude/Cursor/Copilot can query. Headlamp ships an in-app AI Assistant that can also act as an MCP client.
MCP endpoint (use any client)
AI Assistant plugin
Karpenter / KEDA / cert-manager / Kyverno / OpenCost / Trivy
All are cluster operators you install yourself. The question is what the UI does once they're there. Radar detects each and renders integrated views. Headlamp ships per-tool plugins from headlamp-k8s/plugins that you install separately.
Auto-detected, no plugin install
Plugin per integration
Collaboration
Slack / PagerDuty / MS Teams alerts
Cloud
Resource annotations visible to teammates
Cloud
Webhooks
Free for unlimited teammates
Yes - it's the whole tool
Compliance
SOC 2 Type 2 (hosted)
Cloud
N/A - self-host
Self-host / on-prem
OSS · Cloud Enterprise
Helm chart

Based on Headlamp v0.41 (March 2026) and the kubernetes-sigs/headlamp + headlamp-k8s/plugins repos. Something wrong? email us.

Concrete scenarios

What each tool does for the jobs that fill your week.

2am production incident

Radar
Headlamp

Open Headlamp, pick the cluster, browse to the workload, see current state. If the trigger event already rolled off K8s' ~1h retention, reconstruct from logs.

Radar

Click the Radar link in the PagerDuty alert (Cloud). Rewind the timeline to before the page fired. Paste URL into the incident channel.

Compare staging to prod

Radar
Headlamp

Side-by-side compare panel. Pick two clusters, see resources next to each other.

Radar

Drift-diff tool across the whole environment, not one resource type at a time.

Auditor asks what changed last quarter

Radar
Headlamp

No persisted audit. Reconstruct from git, kubectl history, external SIEM if you ship there.

Radar

Filter audit log by namespace + date range. Export CSV.

Onboard a new engineer with Okta SAML

Radar
Headlamp

Configure OIDC (Keycloak / Dex / OpenUnison), set up the cluster ServiceAccount, walk them through the URL.

Radar

SAML SSO into Radar Cloud (SCIM on Enterprise). First login, they're in.

Where's the cluster spend going?

Radar
Headlamp

Install the OpenCost Headlamp plugin. Configure it.

Radar

Cost view is on by default, tied to the workloads.

Why did ArgoCD just sync?

Radar
Headlamp

Open ArgoCD UI separately - the official plugin set is Flux-only.

Radar

Sync status, source revision, and diff inline on the resource - same for ArgoCD or Flux.

Vendor-neutral governance for procurement

Headlamp
Headlamp

kubernetes-sigs sub-project, CNCF Sandbox. Easier procurement story.

Radar

Apache 2.0 OSS with no-relicense pledge - solid, but a single vendor underneath.

Solo engineer, laptop only, no cluster footprint

Tie
Headlamp

brew install headlamp, open, point at kubeconfig.

Radar

brew install radar, kubectl radar, browser opens.

FAQ

Questions we get about Headlamp.

Is Radar trying to replace Headlamp?
No. Headlamp is a plugin-extensible dashboard with kubernetes-sigs governance. Radar is what we built when we wanted real-time topology, persistent timeline, and traffic visualization at the architecture level, not via a plugin shim.
Same Apache 2.0 license. Why not just write a Headlamp plugin?
Plugins are bounded by what the plugin API exposes, with their own performance ceiling and lifecycle to maintain. Some things - real-time topology, server-pushed timeline updates - aren't shaped like a plugin. For everything else: contributing a feature into Radar core is less work than writing and shipping a Headlamp plugin for the same thing.
Does Radar have a desktop app like Headlamp?
Radar is a single Go binary with the UI embedded. brew install, run kubectl radar, browser opens against your current kubeconfig in under 15 seconds. No Electron, no per-OS installer to maintain. Headlamp's per-OS desktop installers (Brew/WinGet/Choco) are smoother for the laptop-only solo case.
Headlamp has a Flux plugin. What does Radar add?
Headlamp's official Flux plugin is solid; ArgoCD isn't part of the official set. Radar correlates resources to both ArgoCD and Flux out of the box. The deeper difference: Radar puts the sync state next to the resources it produced and ties it into the timeline, rather than rendering the GitOps CRDs as their own section.
Headlamp has an AI Assistant. Is that the same as Radar's MCP support?
Different shapes. Headlamp's AI Assistant is an in-app chat panel that can also act as an MCP client. Radar exposes the cluster as an MCP endpoint - your existing Claude / Cursor / Copilot queries it directly with no in-app chat UI.
Is Headlamp actively maintained?
Yes. Monthly release cadence, v0.41 in March 2026, 126 total releases, kubernetes-sigs governance, adopted by Microsoft, Oracle, and Swisscom. The pre-1.0 versioning is a Kubernetes-ecosystem norm, not a maintenance signal.

Try Radar OSS in 30 seconds.

Single Go binary, Apache 2.0. Or use hosted Radar Cloud free for 3 clusters.

Apache 2.0 · Self-host forever · Cloud for fleet, alerts, SSO