Radar

Radar is an open-source Kubernetes control and visibility interface for platform engineering teams, with views designed for both operators and AI agents. It is a strong fit when topology, local deployment and Kubernetes integrations matter more than long event history or low per-cluster Cloud pricing.

Overview

Radar brings live topology, resource inspection, Helm, GitOps, image filesystems, audits and MCP/AI functions into one interface. It runs as a single binary against a kubeconfig or in-cluster through Helm, while desktop applications cover macOS, Linux and Windows. That makes it practical for engineers who move between a workstation, a cluster and a shared web service.

The open-source edition keeps cluster data contained: it sends no cluster data out and makes one anonymous version check that can be blocked. Cloud streams viewed resources, events and logs through its control plane without storing them. Cloud is therefore convenient for hosted investigations, but teams with strict data policies should decide whether that transit model is acceptable.

Key features

Topology and integrations

The live graph links Deployments, Services and Ingresses, showing traffic, TLS health and cross-namespace dependencies. That relationship-focused view is more useful for tracing service impact than isolated resource pages. Built-in integrations include Argo, Flux, Cilium, Istio, Karpenter, KEDA, Crossplane, cert-manager, OpenCost, Kubecost, Prometheus, Trivy and Velero, with no plugin installation.

Investigations and notifications

Radar Cloud adds AI root-cause investigations with evidence, plus per-cluster routing to Slack or any webhook. The allowance is metered, so investigation-heavy teams need to watch usage. Radar does not provide Kubernetes logs or distributed tracing as dedicated capabilities, which limits its role as a complete observability replacement.

Security and operations

SOC 2 Type 2 applies to every Cloud plan. Enterprise and Custom add SAML/OIDC single sign-on and SCIM 2.0. OSS support is through GitHub and Discord; Enterprise provides priority support, while Custom includes a dedicated customer success manager and phone support.

Pricing

Free options

Radar OSS costs nothing, billed Free forever. It includes unlimited clusters and nodes, full topology and resource views, Helm, image filesystem, audit and MCP/AI features. This is the best route for teams that can operate the software themselves and require local data handling, but support remains community-based.

Cloud Free costs nothing, billed $0 forever. It allows up to three clusters, unlimited users, 100 hosted investigations per month and 24-hour event-timeline retention. It is suitable for evaluation or small estates; the cluster cap and one-day history quickly constrain broader operations.

Paid Cloud plans

Team starts at $149 a month, billed $149 per cluster / month. It provides unlimited clusters and users, 100 investigations per cluster each month, three-day timeline retention and 30-day audit retention. Extra investigations cost $1 each if enabled. Team fits production groups that need hosted access without enterprise identity controls, but per-cluster billing becomes significant as estates grow.

Enterprise costs $299 a month, billed $299 per cluster / month. It adds SAML/OIDC SSO, SCIM 2.0, 200 investigations per cluster monthly, 30-day timeline retention and one-year audit retention; additional investigations are $1 each if enabled. Choose it when identity lifecycle management and longer audit history justify the premium.

Custom uses an annual contract with volume pricing. It adds a custom identity provider and data residency, custom retention, air-gapped deployment and a 99.9% uptime SLA. This is the plan for regulated or large environments that cannot use standard Cloud terms.

A 14-day trial applies to the paid offering, giving prospective buyers a short window to validate cluster coverage and investigation consumption before renewal commitments.

Platforms

Radar supports API, Linux, macOS, self-hosted, web and Windows deployments. The single-binary and Helm options suit Kubernetes-native operations, while desktop apps provide local access without limiting teams to a browser.

Who it's for

Radar is aimed at platform engineering teams that need centralized Kubernetes visibility for people and AI agents. OSS is the better choice for organizations prioritizing data containment and unlimited cluster coverage. Cloud suits teams wanting managed investigations, alert routing and shared access. It is less suitable as a standalone observability stack when logs, traces or long historical timelines are requirements.

Pros and cons

Pros

• Live topology exposes traffic, TLS and cross-namespace relationships.
• Broad built-in Kubernetes integrations avoid plugin maintenance.
• OSS supports unlimited clusters and keeps cluster data local.
• Cloud offers evidence-based AI investigations and Slack/webhook alerts.
• Enterprise identity controls and one-year audit retention address larger teams.

Cons

• Team and Enterprise charge per cluster, which can outpace seat-based tools.
• Cloud event retention is only 24 hours on Free, three days on Team and 30 days on Enterprise.
• No dedicated Kubernetes logs or distributed tracing.
• OSS relies on GitHub and Discord rather than priority support.

Alternatives

Choose Kubernetes Monitoring when you want a category-wide comparison before selecting a tool.

Middleware is a better fit for teams needing broader application telemetry, with a free tier covering up to 100GB of data, 1,000 RUM sessions and 20,000 synthetic checks. Netdata suits lightweight infrastructure monitoring, while its Community plan is limited to five active connected nodes. Dynatrace Kubernetes Monitoring is the alternative for paid, usage-based Kubernetes monitoring at $0.002 per pod hour.

Grafana Cloud Application Observability fits teams wanting a free allowance of 2,232 host hours and pay-as-you-go Pro pricing. Checkmk is preferable for traditional infrastructure coverage, with a Community edition supporting about 100 hosts and more than 2,000 integrations. Datadog Kubernetes Monitoring offers container-hour billing, while New Relic Kubernetes Monitoring combines annual Pro commitments with 100GB of free ingest. VictoriaMetrics is the alternative when an open-source metrics platform is the priority.

Verdict

Choose Radar for a Kubernetes-focused operational map with local OSS deployment, deep integrations and optional AI investigations. Look elsewhere if your team needs built-in logs and tracing, long Cloud retention or predictable pricing that does not rise with every cluster.