K8s Status

Monitor Kubernetes nodes, pods, and deployments from Noctalia, with a /kube launcher for common actions.

Plugin

Field	Value
ID	davemhammer/k8s-status
Entries	Bar widget: status; panel: manager; service: service; launcher: kube
Launcher Prefix	/kube

Requirements

Install these on PATH (declared in plugin.toml dependencies):

• kubectl — cluster queries and actions (or set kubectl_bin)
• less — pager for describe/logs in the terminal

Optional (not declared; used only if present):

• k9s — open-k9s action

Cluster access uses your kubeconfig (passed to kubectl as --kubeconfig; the plugin does not parse the file itself).

Usage

Add the status bar widget (davemhammer/k8s-status:status). Click for the panel; right-click requests a refresh.

Panel tabs: Nodes, Pods, Deployments, Namespaces. Select a row for describe / logs / shell / restart actions (where applicable).

Launcher examples:

• /kube — categories (pods, problems, nodes, …)
• /kube pods nginx — filter pods
• /kube problems — problem pods only

noctalia msg panel-toggle davemhammer/k8s-status:manager

Settings

Setting	Type	Default	Description
kubeconfig	file	~/.kube/config	Path passed to kubectl as --kubeconfig.
context	string	(empty)	Context name; empty uses current-context.
namespace	string	(empty)	Limit pods/deployments; empty = all namespaces.
refresh_interval	int	15	Poll interval in seconds.
problems_only	bool	false	Prefer problem pods in the panel list.
notify_on_not_ready	bool	true	Notify when a node becomes NotReady.
kubectl_bin	string	kubectl	kubectl command or absolute path.
show_counts	bool (widget)	true	Show ready nodes / problem pods on the bar.
ok_color	select (widget)	tertiary	Bar color when cluster looks healthy.
warn_color	select (widget)	error	Bar color when there are problems.

IPC

noctalia msg panel-toggle davemhammer/k8s-status:manager
noctalia msg plugin davemhammer/k8s-status:service all refresh

Notes

• Shells out to kubectl and less (and optionally a terminal via runInTerminal for logs/shell/k9s).
• Refresh: nodes, deployments, and namespaces use compact jsonpath queries; pods use kubectl get pods table output for ready/restart columns.
• Does not modify cluster state unless you run restart/delete-style actions from the panel or launcher.
• Network: only via kubectl to the API server from your kubeconfig. No cluster credentials are written into the plugin tree.