Notes from the team building Radar.
Deep dives on Kubernetes visibility, multi-cluster architecture, incident patterns, and the engineering behind the product.
More from the team
Why We Priced Radar Per Cluster, Not Per Seat
Per-seat pricing punishes the thing a visibility tool is for: more eyes. Here's the five pricing models we compared, why we picked per-cluster, and where it's a worse deal.
Radar OSS or Radar Cloud? An Honest Take on When to Use Each
The most common question in our GitHub issues: do I need Radar, or is OSS enough? Here's the honest answer, with the five questions that actually decide it.
Multi-Cluster Topology: Cross-Cluster Service Maps That Don't Hairball
Cross-cluster service topology is hard because Kubernetes itself has no multi-cluster graph. Here's how Radar builds one without turning it into a hairball.
Your Cluster Events Are Mostly Noise. Here's How We Filter Them.
Every team has a #k8s-alerts Slack channel no one reads. Here's why naive event alerting fails and how Radar correlates, suppresses, and routes the 2% that matters.
SSO, SCIM, and Scoped RBAC: Team Access Without YAML Wrestling
A contractor needs read-only access to two staging namespaces by Monday. Without Radar, that ticket bounces for a week. Here's how scoped RBAC fixes it.
Persistent Event Timeline: Debugging What Happened Last Tuesday
Kubernetes events vanish after an hour. Radar keeps them for 30 days. Here is how a persistent event timeline changes incident response.
Fleet-Scale Informers: Watching Thousands of Watch Streams Without Melting
SharedInformer is great until you need a thousand of them. How Radar's backend keeps a live server-side replica of every connected cluster without melting.
The Radar Agent: Outbound-Only, Scoped, and 32MB
How the Radar agent works: outbound-only TLS, read-only RBAC by default, ~32MB RSS. The five questions every platform team asks, answered.
What We're Building in 2026: Radar OSS and Radar Cloud
Two shipping dates, one product line. Radar open source ships January 2026, Radar follows in February. Here's the scope, the split, and what won't make it into v1.
The Fleet Visibility Gap: Why Teams With 5+ Clusters Hit a Wall
Every tool that worked at 2 clusters breaks at 8. kubectl, Lens, k9s, Headlamp are all single-cluster-at-a-time. Here's where the wall is and what it looks like.
Get posts in your inbox.
Monthly, not noisy. Engineering deep dives + launch notes. Unsubscribe any time.
Newsletter form is placeholder — wire to a real provider before launch.
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