Your cluster, in plain sight.

Unhealthy namespaces sort first and healthy ones collapse to a single line - the whole cluster's problems arrive on one screen.

The screen opens on the reason, not the metadata - last termination on top, with containers, resource pressure and events in one view.

Walk Certificate to CertificateRequest to Order to Challenge, with the API's exact failure reason preserved at every step.

Warnings, container restarts and rollout revisions share one newest-first clock, so cause and consequence stay together.

Jump to any kind or object through one fuzzy palette, with short aliases for the daily resources you reach most.

Switch namespace with live pod health in view. The palette opens on the last other namespace, so n + ↵ toggles back.

Change context after a reachability check, with each context's namespace and your last other context remembered.

The latest terminal failure sits above scheduled and manual run history, with the controller's own reason and a direct path to the exact Job.

Each retained retry keeps its result, exit code, duration and node together; rerun stages a safe copy or an explicit replacement.

The deployed chart sits beside the newest version in your local repo cache, so current and outdated releases stay distinct without the helm binary installed.

Read values with v and walk every revision with h, still without Helm installed. Only rollback with R invokes the Helm CLI.

Every host and path resolves to the backend that actually answers - endpoint health and TLS included.

See every match, weighted backend and parent Gateway together, with live endpoint health behind each route.

Each source keeps the reconcilers it drives underneath it, with failed, suspended, reconciling and source-behind states kept distinct.

Argo CD Applications keep sync and workload health separate, with the sickest managed resource's own message directly under a degraded row.

Press + or − to start from the next replica count. If an HPA owns replicas, kute warns that it will overwrite the change before you apply.

Pick a tag from this workload's rollout history or one already running elsewhere, instead of typing it from memory.

Edit requests and limits beside the container's usage, with the exact change shown before you apply.

Start a port-forward and keep browsing. The header tracks it globally; the Forwards list shows its local port, resolved target, and status.

Outside production-tagged contexts, delete stays inline: the table remains visible beside the exact command and a simple y/N confirmation.

Delete escalates to typing the full name only in contexts you explicitly mark as production - never because their names merely look important.