Replacing OpenLens: why I kept K9s, Headlamp, and Kubenav

OpenLens is no longer maintained. The last update was months ago, and the community has moved on. If OpenLens was your Kubernetes UI, you need a replacement.

I tested 13 tools across my production and dev clusters. I kept three, and each one has a different job.

Why OpenLens stopped being an option

Lens started as an open-source project but gradually moved toward a commercial model. OpenLens, the community fork, kept the open-source version alive until it didn't. Development stalled, bugs went unfixed, and security patches stopped coming.

A new fork called Freelens has picked up where OpenLens left off. It is the closest thing to OpenLens. I still wanted to see what else was out there before committing.

What matters during an incident

A dashboard is only as good as its behavior during a live cluster incident.

A dashboard that chokes on ten thousand pod logs will waste precious minutes when nodes are flapping. Pick a tool that boots instantly, exposes crash backtraces without round-trips, and works over minimal terminal or SSH bandwidth.

Not every role needs all three. A teammate checking pod status in a browser has different needs than someone debugging over SSH.

The decision: three tools for three roles

K9s is my daily driver

K9s is a terminal UI for Kubernetes. I use it for quick cluster checks and debugging. Nothing beats the speed of terminal-based navigation.

It is extremely lightweight, uses few resources, and works well in SSH sessions and remote environments.

Navigation is keyboard-driven. You get real-time resource monitoring, multiple clusters and contexts, built-in resource editing, and a plugin system.

The cost is a steeper learning curve for terminal-averse users. Graphical visualization is also limited compared to GUI tools. I accept both for the speed.

# macOS
brew install derailed/k9s/k9s

# Linux (via snap)
sudo snap install k9s

# Windows (via scoop)
scoop install k9s

# Or download from GitHub releases
# https://github.com/derailed/k9s/releases

Headlamp serves teammates who want a web UI

Not everyone wants to live in a terminal. For team members who prefer a web UI, I deploy Headlamp in-cluster.

Headlamp started at Kinvolk, which Microsoft acquired in 2021, and is now a Kubernetes SIG UI project. That backing gives me confidence in its longevity.

It has a clean web interface, multi-cluster management, RBAC-based access control, and real-time updates. It can run in-cluster or locally, and its resource visualization is excellent. Of the web options, it is the most production-ready.

The downsides are real but small for this role. It is relatively new, so the ecosystem is still developing. It also has fewer advanced features than Lens.

# Helm install (in-cluster)
helm repo add headlamp https://kubernetes-sigs.github.io/headlamp/
helm install headlamp headlamp/headlamp

# Desktop app (macOS)
brew install --cask headlamp

# Desktop app (Linux)
flatpak install flathub io.kinvolk.Headlamp

Kubenav covers weekend on-call from my phone

Kubenav is the only tool I found with iOS and Android apps. It also runs on desktop and web. I keep it on my phone for weekend on-call monitoring.

Being able to check pod status from anywhere is useful. It supports multiple clusters, Prometheus metrics, terminal access to containers, and RBAC. It also works offline with cached data.

It has fewer advanced features than desktop-focused tools, and its community is smaller. For a status check from a phone, that trade is fine.

What I tried and did not adopt

Kubernetes dashboards fall into three groups: terminal UIs, desktop apps, and web UIs.

The other tools are each good at something; I just didn't need them.

If your workflow differs from mine, this map covers the rest:

Not tested yet: Radar

After this post went out, Radar's CEO messaged me to pitch Radar. I have not run it on my clusters, so it is not in the tables above.

From its README: it ships as a single Go binary under Apache 2.0. It runs from your laptop against the Kubernetes API and serves a web UI. It can also be deployed in-cluster with Helm for shared team access. It has a built-in MCP server for AI agents. It had about 3,760 GitHub stars in October 2026.

If you try it before I do, tell me how it holds up during an incident.

Migration path from OpenLens

If you are on OpenLens today, here is the order I recommend:

1. Try Freelens first. It is the closest thing to OpenLens and needs minimal adjustment.
2. Learn K9s on the side. Even if you stick with a GUI, K9s will make you faster.
3. Evaluate Headlamp for team use. If you need shared access, it is the most production-ready web option.

Full comparison table

Lens and the 12 alternatives side by side, for reference.

Final thoughts

I expect more dashboards to target specific niches rather than trying to do everything. That suits a split setup: one tool per role.

What replaced OpenLens on your team? Let me know in the comments.

Resources

• Freelens
• K9s and K9s releases
• Headlamp
• Kubenav
• Radar (not tested yet)