Built for the work that eats your day.
Incidents. Rollouts. Drift. Audits. These are the stories we've heard from platform teams — and the ones Radar was designed to end.
Stop playing kubectl roulette at 3am.
A pod is crashing. You don't remember which cluster, which namespace, or which alert fired. You open three terminal tabs, SSH through a jump host, try to find the right kubeconfig, and by the time you're in — the pod has restarted and the logs are gone.
Radar gives your on-call a single search bar across every cluster. Paste a pod name. Jump to the timeline. Rewind to before the crash. Share a link with the incident channel.
- MTTR drops from 20+ minutes to under 5
- On-call engineers stop needing to remember which cluster a service runs in
- Post-incident reviews have actual evidence, not 'I think the pod crashed around 2:15'
- New hires ramp to productive on-call in days, not months
Manage a fleet, not a cluster.
You run dozens of clusters across staging, prod, regions, and maybe a few legacy environments you're afraid to touch. Each one has its own kubeconfig, its own dashboard, its own rules. Comparing them requires scripts.
Radar shows every cluster in one view. Drift detection highlights where staging and prod diverged. Cross-cluster search finds every instance of a Deployment in one query. You see the whole fleet as a fleet.
- Drift between environments gets caught before it causes outages
- Workload inventory is continuously accurate — no more spreadsheets
- Region-by-region rollouts are visible in real time
- Unified cost view across EKS, GKE, AKS, and on-prem
See what ArgoCD actually did.
ArgoCD says 'Synced and Healthy.' But Pod is in CrashLoopBackOff. Which commit caused this? What does the live state actually look like? The ArgoCD UI shows the intent — not the reality.
Radar shows ArgoCD and Flux sync state next to the live Kubernetes resources they produced. See the commit, the diff, the resulting workload, and the events that followed — in one timeline.
- Correlate a bad deploy with the exact commit that caused it
- See the delta between 'ArgoCD synced' and 'cluster is actually healthy'
- Audit every GitOps change with a persistent timeline
- Catch drift between Git and live state
Helm release visibility that doesn't require helm list.
You have 200 Helm releases across 15 clusters. Some are on chart v3.2, some on v4.1, some on a fork someone made in 2022. Nobody knows which version is where. Rollback requires finding the right kubeconfig first.
Radar inventories every Helm release across every cluster. See chart versions, revision history, rendered values, and diffs between revisions. Roll back from the UI with a single click.
- Catch chart-version drift across environments
- Roll back a bad release in one click, from any cluster
- Audit changes to values files over time
- Trace a production issue to the exact values-file change that caused it
Answer the auditor's question without grepping logs.
Your SOC 2 auditor asks: 'Show me every change to the production payments namespace in Q3.' You have Kubernetes events — if you were recording them. You weren't. You have kubectl history — but only on the machines that ran the commands.
Radar persists every Kubernetes event, every RBAC change, every Helm revision, and every user action for up to a year. Export to your SIEM. Hand the auditor a filtered timeline.
- SOC 2 audit evidence in one export
- Audit logs that survive cluster restarts and node cycles
- Namespace-scoped RBAC that maps to your identity provider
- Proof of who did what, when, and across which cluster
Know which workload is burning your cloud bill.
Your AWS bill went up 18% last month. Your finance team wants to know why. You have AWS Cost Explorer, which tells you EKS cost you more. Great — but which workload, in which cluster, is responsible?
Radar integrates OpenCost across every cluster. Attribute cost to workload, namespace, label, or team. Compare cluster-by-cluster efficiency. Identify the 20% of workloads responsible for 80% of the spend.
- Workload-level cost attribution without deploying another agent
- Team-by-team chargeback using existing Kubernetes labels
- Idle-resource detection with reclaim suggestions
- Cross-cluster efficiency benchmarking
Bring your first cluster online in 60 seconds.
Install the Helm chart, paste a token, see your cluster. No credit card required.
Apache 2.0 OSS · Unlimited clusters self-hosted · Hosted free tier for up to 3 clusters