Glowpod in the terminal

glowpod term is Glowpod as a terminal app: the same Glance view of your cluster, the same plain-language answers and the same live resource lists as the web dashboard, with keys instead of clicks. It's part of the glowpod program, so there's nothing extra to install. It runs wherever your kubeconfig works, including over SSH.

Glance in the terminal: CPU, memory, pods, storage and network gauges, a pod that keeps restarting under Needs attention, and recent events

Install

macOS and Linux, with Homebrew:

Terminal

brew install glowpod-dev/tap/glowpod

Windows, with Scoop:

PowerShell

scoop bucket add glowpod https://github.com/glowpod-dev/scoop-bucket
scoop install glowpod

Or download glowpod for your system (Linux, macOS or Windows, Intel or ARM) from the latest release. On Windows, use Windows Terminal: older consoles draw the gauges poorly.

Start it

Terminal

glowpod term                   # Glance, across all namespaces
glowpod term -n shop           # in one namespace
glowpod term -n shop deploy    # start on a table: pods, deploy, svc, any CRD
glowpod term -context staging  # another kubeconfig context

It reads your kubeconfig the way kubectl does (-kubeconfig, -context, $KUBECONFIG, then ~/.kube/config) and acts as that user. Everything you see and do is what your RBAC allows.

Glance and Deep

Glance opens first: the cluster's CPU, memory, pod, storage and network gauges, the things that need attention in plain words ("Keeps restarting (CrashLoopBackOff)") with what to check next, and recent events.

Press Tab for Deep: any resource as a live table with the API server's own columns, the same ones kubectl get shows. CRDs work too. w adds the wide columns, / filters as you type (/-l app=web filters by label), and n switches namespace.

Deep: the shop namespace's pods, with a pod in CrashLoopBackOff marked in red

Logs, events and YAML

Open a pod with Enter, or press l on it, for its live logs. As in the web dashboard, a line's own level decides its color: warnings orange, errors red. 1, 2 and 3 switch to Events and YAML, p shows the previous run after a crash, t adds timestamps and G follows new lines again.

Live logs of a pod, with warning lines in orange and error lines in red

The command palette

: (or ctrl+k) opens one prompt for everything: any resource by name, singular, kind or short name (:po, :deploy, :svc, any CRD), a namespace (:ns monitoring, :ns all), screens (:glance, :help) and actions on the selected row. Tab completes and Enter runs it.

The command palette, suggesting deployments, daemonsets and other kinds for "dep"

Keys

? shows every key for the screen you're on.

Key	Does
`Tab`	Glance ↔ Deep
`:` or `ctrl+k`	Command palette
`/`	Filter the list (`-l key=value` for labels)
`n`	Pick a namespace
`↑` `↓` or `j` `k`, `g` `G`	Move, top, bottom
`Enter`	Open
`l`	Logs
`y`	YAML
`w`	Wide columns
`s`	Scale a workload, or a shell in a pod
`r`	Restart a Deployment, StatefulSet or DaemonSet (on Glance: refresh)
`ctrl+d`	Delete (type the name to confirm)
`Esc` or `q`	Back; `q` on the first screen quits
`ctrl+c`	Quit

Scale, restart, delete and shell only appear when the cluster says you may do them. If your rights change while it's open, a refused action says so in plain words.

Help, listing the keys for the pod list and the keys that work everywhere

Coming next

Revealing Secret values, editing YAML, port-forward, a light theme, and signing in to your team's in-cluster Glowpod (SSO, LDAP or a token) instead of using a kubeconfig.

1Install

2Start it