Your single pane of glass into Kubernetes

What you can do with it

• See what needs attention

  The Overview lists every object with a problem, errors first and then the newest, with the reason in plain words. An object's page says what is wrong and checks for the likely cause.

• Fix it where it shows

  Revert the change made just before a problem, give a container more memory, change an image, roll back, run a CronJob again or sync an Argo CD application, from the problem itself.

• Every kind, one list

  Built-in kinds and custom resources get the same live list and detail page: status from the server, label filters, sorting, bulk actions and k9s keys. On a wide screen an object opens beside its list.

• See how it connects

  The map lays out Ingresses, Services, workloads, pods and what they mount, and marks the chain that leads to a broken object, such as an Ingress whose Service has no ready endpoints.

• Read the logs

  Follow one pod's logs, or a Deployment's pods merged in time order. Filter by text, regex or level, expand JSON, and read the run before a restart.

• Change it, see what changed

  Edit YAML with a dry run and a diff, or a Secret's or ConfigMap's values decoded, then scale, restart, roll back, cordon and drain. KubeGlass records each change field by field, with who made it, and can undo one. From the first launch the history also holds the rollouts and Helm revisions the cluster kept from before.

• Open a shell

  Pod and node shells in split panes, each in its cluster's color, with statuses such as CrashLoopBackOff colored. Broadcast to several clusters and record sessions. A debug container gets into images without a shell. Browse a container's files, download them and upload files of up to 256 MiB. Start with --local-shell for a shell on your own computer.

• Check access

  What you or any user, group or ServiceAccount may do, who a binding lets do something, and a form that writes Roles and bindings.

• Applications, Helm and GitOps

  Workloads grouped by Helm release, Argo CD application or labels, each with its own page. Helm releases and a chart catalog, Argo CD, Flux and Argo Rollouts.

• Work across clusters

  Every cluster in one list with its problems and its own color, lists and search across several clusters, production and read-only marks, a diff between clusters, drift detection and an upgrade check.

• Group by team

  A project is the namespaces that share one label, such as a team's dev, staging and prod, with their problems, applications, map and access on one page. Projects made in Headlamp show up too.

• CPU and memory history

  Pod, workload and node pages chart the last day of CPU and memory against requests and limits, from Prometheus or, without it, from metrics-server samples KubeGlass keeps.

• Monitoring and security

  Pods that keep restarting, Alertmanager alerts and silences, rightsizing, PromQL, best-practice checks, certificates, vulnerabilities, images and policy violations.

• Hear when something breaks

  The bell lists the problems that started since you last looked. Turn on a toast, a chime or a desktop notification, with quiet hours.

• Add your own

  One YAML file adds actions, keys, sidebar links, Grafana panels, list columns and palette names. Import your k9s plugins, hotkeys and aliases into it.

• Ask from your editor

  The MCP server lets AI clients such as VS Code and Claude Desktop query your clusters through KubeGlass. Its 41 tools only read.

Who it's for

• Developers

  See your Deployment's rollout and pods, read the logs of all its pods together, open a shell, and edit YAML with a dry run and a diff before you apply it.

  Applications

• On call

  The Overview lists what is failing, errors first and then the newest, with the reason and the change made just before it began. Revert that change, roll back or restart from the problem, and see who changed what in the Timeline.

  The Overview

• Platform teams

  Every cluster in one list with its problems and its own color. Mark production and read-only clusters, compare two clusters, detect drift on a schedule, check upgrade readiness, and see Helm, Argo CD and Flux together.

  Clusters

• Security

  See what any user, group or ServiceAccount may do, who may do something, and the role that grants it. Checks, certificates, vulnerabilities, policy violations and a log of every change.

  Access

How it compares

KubeGlass next to the Kubernetes UIs people most often weigh it against.

Checked on 5 October 2026 against each tool's own docs, release notes and source. Sources and details

Kubernetes Dashboard was archived on 21 January 2026. Moving from Kubernetes Dashboard to KubeGlass

Install

brew install kubeglass/tap/kubeglass
kubeglass

docker run --rm -p 8090:8090 \
  -v ~/.kube/config:/home/nonroot/.kube/config:ro \
  -v kubeglass-data:/home/nonroot/.kubeglass/data \
  ghcr.io/kubeglass/kubeglass:latest

helm install kubeglass oci://ghcr.io/kubeglass/charts/kubeglass \
  --namespace kubeglass --create-namespace \
  --set config.authMode=local
kubectl -n kubeglass port-forward svc/kubeglass 8090:8090

git clone https://github.com/kubeglass/kubeglass.git
cd kubeglass
make build
./bin/kubeglass

Then open localhost:8090. On your own machine KubeGlass listens only on 127.0.0.1, because anyone who reaches it gets your cluster access. Before you share a deployment, read Running for a team. All the options are in the installation guide.

Questions