DEV Community

Cover image for K9s — The Terminal UI That Changes How You Do Kubernetes
Shiva Shanmugam
Shiva Shanmugam

Posted on

K9s — The Terminal UI That Changes How You Do Kubernetes

#kubernetes #devops #cli #productivity

Stop memorizing endless kubectl commands. K9s gives you a real-time, interactive terminal dashboard for managing every resource in your cluster — with keyboard shortcuts for everything.

Table of Contents

Why kubectl Alone Isn't Enough

Anyone who has spent serious time with Kubernetes knows the pain: debugging a crashing pod means running four or five different commands just to understand what's happening. You're constantly context-switching between terminal windows, squinting at JSON blobs, and copy-pasting pod names.

# The old way — just to debug ONE crashing pod
kubectl get pods -n production
kubectl describe pod api-gateway-7d9f8b-xkp2q -n production
kubectl logs api-gateway-7d9f8b-xkp2q -n production --previous
kubectl exec -it api-gateway-7d9f8b-xkp2q -n production -- /bin/sh
kubectl port-forward api-gateway-7d9f8b-xkp2q 8080:80 -n production
Enter fullscreen mode Exit fullscreen mode

That's five commands, three copy-pastes, and a lot of cognitive overhead — just for one pod.

kubectl vs K9s Side by Side

Task ⚡ K9s kubectl
View all pods :pod then Enter kubectl get pods -A
Stream pod logs Select + l kubectl logs -f <pod-name>
Exec into pod shell Select + s kubectl exec -it <pod> -- /bin/sh
Port-forward Select + Shift+F kubectl port-forward <pod> 8080:80
Delete a resource Select + Ctrl+D kubectl delete pod <name>
Edit resource YAML Select + e kubectl edit pod <name>
Switch namespace Ctrl+A for all namespaces -n flag on every command
Real-time CPU/Mem ✅ Built-in pulse view ❌ Requires separate tooling
Rolling restart Select + Ctrl+L kubectl rollout restart deploy/<name>
Decode secrets Select + x kubectl get secret <name> -o jsonpath=...

What is K9s?

K9s is an open-source, terminal-based UI built by Fernand Galiana that wraps the entire Kubernetes API in a fast, keyboard-driven interface.

It's not a GUI — it lives in your terminal, it's blazing fast, and once you learn the shortcuts, you'll never want to go back to plain kubectl.

💡 Think of K9s as htop for Kubernetes — it continuously watches your cluster state and updates in real time.

 ____  __.________
|    |/ _/   __   \______
|      < \____    /  ___/
|    |  \   /    /\___ \
|____|__ \ /____//____  >
        \/            \/

K9s — Kubernetes CLI To Manage Your Clusters In Style!
Enter fullscreen mode Exit fullscreen mode

Key facts:

  • ⭐ 30,000+ GitHub stars
  • 🐕 Named after the "canine" unit — it "watches" your cluster
  • 🔄 Real-time polling — no manual refresh needed
  • 🔐 Fully RBAC-aware — respects your cluster permissions
  • 🧩 Supports every Kubernetes resource including CRDs

Installation

K9s runs on Linux, macOS, and Windows. Pick your method:

macOS — Homebrew 

brew install derailed/k9s/k9s
Enter fullscreen mode Exit fullscreen mode

Linux 

# Via webinstall
curl -sS https://webinstall.dev/k9s | bash

# Via Snap
sudo snap install k9s

# Via apt (Ubuntu/Debian)
sudo apt install k9s

# Via binary — download from GitHub releases
wget https://github.com/derailed/k9s/releases/latest/download/k9s_Linux_amd64.tar.gz
tar -xzf k9s_Linux_amd64.tar.gz
sudo mv k9s /usr/local/bin/
Enter fullscreen mode Exit fullscreen mode

Windows 

# Chocolatey
choco install k9s

# Scoop
scoop install k9s
Enter fullscreen mode Exit fullscreen mode

Docker 

docker run --rm -it \
  -v ~/.kube/config:/root/.kube/config \
  quay.io/derailed/k9s
Enter fullscreen mode Exit fullscreen mode

Verify Installation 

k9s version
k9s   # Launch!
Enter fullscreen mode Exit fullscreen mode

K9s auto-detects your ~/.kube/config on launch.

Launch Flags

Command Description
k9s Launch with current kubeconfig context
k9s -n kube-system Launch in a specific namespace
k9s --context prod Launch in a specific kube context
k9s --readonly Launch in read-only mode (safe for production)
k9s --kubeconfig /path Use a custom kubeconfig file
k9s --all-namespaces Start with all namespaces visible

⚠️ Production Safety Tip: Always use k9s --readonly when connecting to production clusters. This disables all mutating operations and prevents accidental deletes or edits.

Core Features

🔴 Real-Time Cluster State

K9s continuously polls your cluster and refreshes the view automatically. Watch pods crash and recover without hitting a single key.

⌨️ Fully Keyboard Driven

Every action — logs, exec, describe, delete, port-forward — is one or two keystrokes away. Zero mouse needed.

📋 Universal Resource Access

Type :pod, :svc, :deploy, :node or any CRD name. K9s handles every Kubernetes resource type including custom ones.

📊 CPU & Memory Metrics

Built-in pulse view shows live resource usage per pod and node. Color-coded thresholds alert you before things go down.

🔍 Fuzzy Search & Filter

Press / to filter any resource list instantly. Supports regex patterns for precise searching across thousands of resources.

🎨 Skins & Plugins

Customize colors, create your own keyboard shortcuts, and extend K9s with community or custom plugins.

🔐 RBAC-Aware

K9s respects your cluster's RBAC configuration. Users only see what they're permitted to see — no special access needed.

📦 Multi-Context / Multi-Cluster

Switch between clusters and contexts instantly without restarting. Perfect for managing dev, staging, and production clusters.

🔖 Built-in Benchmarking

Load test a service endpoint directly from the K9s interface using Ctrl+B on any service. No external tool needed.

🔭 X-Ray View

Renders the full dependency tree of any resource — deployments → replicasets → pods → containers — all in one view.

🧹 Popeye Integration

Type :popeye to run a live cluster sanitizer that scans for misconfigurations, deprecated APIs, and security risks.

Complete Shortcut Reference

💡 Vim users: K9s supports j/k for navigation, g/G for top/bottom, and / to search. You're already home.

Global Navigation

These shortcuts work from any view in K9s.

Shortcut Action
: + resource name Go to any resource view (e.g., :pod, :svc, :node)
? Open help / shortcut cheat sheet
Ctrl+A Toggle all namespaces view
Ctrl+C Quit K9s
Esc Go back / exit current view / cancel filter
/ Enter filter mode (fuzzy search current list)
! Toggle error view
Ctrl+U Clear filter
↑ / ↓ or k / j Navigate up/down (vim-style works!)
g / G Jump to top / bottom of list
Ctrl+R Refresh / reload the current view
:ctx Switch Kubernetes context
:ns Switch namespace
:h View command history
Ctrl+E Hide / show the header bar
Ctrl+P Toggle CPU/Mem pulse panel on/off
z Toggle error panel

Pod Management

Enter :pod or :po to access this view.

Shortcut Action
l Stream logs for the selected pod
p Previous logs (like --previous flag — for crashed containers)
s Open a shell (exec) into the selected pod/container
d Describe the selected pod (full metadata + events)
e Edit the pod YAML in your $EDITOR
y View full YAML of the selected resource
Ctrl+K Force kill / delete the selected pod (no confirmation)
Ctrl+D Delete with grace period (safe delete, with confirmation)
Shift+F Open port-forward menu for the pod
f Show all active port-forwards
Ctrl+W Toggle wide output (shows more columns)
Shift+C Sort by CPU usage
Shift+M Sort by memory usage
Shift+N Sort by pod name
Shift+O Sort by container count
Shift+R Sort by restart count
Shift+T Sort by age (time)
Ctrl+L Rolling restart of the pods in the parent deployment
a View containers within the pod
Enter Drill into pod / select container

Log Viewer

Press l on any pod to enter the log viewer.

Shortcut Action
f Toggle full screen for log view
w Toggle line wrap
↑ / ↓ Scroll up/down through logs
Page Up / Page Down Scroll by full page
g Jump to top (oldest logs)
G Jump to bottom (newest / live logs)
/ Search / filter within logs
s Toggle auto-scroll (tail mode on/off)
t Toggle timestamps in log output
p View previous container logs (for restarted containers)
Ctrl+S Save logs to a local file
0–9 Switch between containers within the pod
c Copy log line to clipboard

Services & Networking

Command / Shortcut Action
:svc View all Services
:ing View all Ingresses
:ep View Endpoints
:netpol View NetworkPolicies
:pf Manage all active port-forwards
Shift+F on a Service Open port-forward dialog for the service
d on a Service Describe service (selectors, ports, endpoints)
y on a Service View full Service YAML
e on a Service Edit service configuration
Ctrl+B on a Service Benchmark the service endpoint (load test)
Enter on an Ingress View ingress rules and backend info

Deployments, ReplicaSets & StatefulSets

Command / Shortcut Action
:deploy or :dp View all Deployments
:rs View all ReplicaSets
:sts View all StatefulSets
:ds View all DaemonSets
Ctrl+L on a Deployment Trigger a rolling restart
s on a Deployment Scale — open scale dialog (set replica count)
d Describe the deployment (events, conditions, strategy)
Enter on a Deployment Drill down into pods belonging to deployment
y View full deployment YAML
e Edit deployment YAML inline
Ctrl+D Delete the deployment (with confirmation)

Nodes & Cluster Management

Command / Shortcut Action
:node or :no View all cluster nodes
Shift+C Sort nodes by CPU usage
Shift+M Sort nodes by memory usage
d on a Node Describe node (taints, conditions, capacity, allocatable)
Enter on a Node View all pods running on that node
y on a Node View full node YAML (labels, annotations)
Ctrl+U on a Node Cordon a node (prevent new pod scheduling)
Ctrl+N on a Node Uncordon a node (re-enable scheduling)
Ctrl+D on a Node Drain the node (evict all pods safely)
:cl Cluster info view (API server, version info)
:ev View all cluster events (warnings, errors)

ConfigMaps, Secrets & Storage

Command / Shortcut Action
:cm View all ConfigMaps
:secret View all Secrets
x on a Secret Decode and view base64 secret values
e on a Secret Edit a secret (values auto-encoded on save)
:pvc View PersistentVolumeClaims
:pv View PersistentVolumes
:sc View StorageClasses
d Describe selected resource
Ctrl+D Delete selected resource

RBAC & Access Control

Command Action
:sa View ServiceAccounts
:rb View RoleBindings
:crb View ClusterRoleBindings
:cr View ClusterRoles
:role View Roles (namespaced)
:hpa View HorizontalPodAutoscalers
:vpa View VerticalPodAutoscalers
:po (PodDisruptionBudgets context) View PodDisruptionBudgets

Jobs & CronJobs

Command / Shortcut Action
:job View all Jobs
:cj View all CronJobs
Ctrl+T on a CronJob Trigger a CronJob manually (creates job immediately)
Enter on a Job View pods spawned by this job
d Describe the job (completions, conditions, duration)
Ctrl+D Delete job

Custom Resources (CRDs)

Command / Shortcut Action
:crd View all Custom Resource Definitions
:<crd-short-name> Navigate directly to any CRD (e.g., :vs for VirtualService)
Enter on a CRD List all instances of that custom resource
d Describe a CRD instance
y View the YAML of a CRD instance
e Edit a CRD instance

All Resource Commands at a Glance

Type these with : from anywhere in K9s to jump to any resource view instantly.

Workloads

Command Resource
:pod / :po Pods
:deploy / :dp Deployments
:sts StatefulSets
:ds DaemonSets
:rs ReplicaSets
:job Jobs
:cj CronJobs

Networking

Command Resource
:svc Services
:ing Ingresses
:ep Endpoints
:netpol NetworkPolicies
:pf Active Port Forwards

Cluster

Command Resource
:node / :no Nodes
:ns Namespaces
:ctx Contexts (switch cluster)
:ev Events
:cl Cluster Info

Config & Storage

Command Resource
:cm ConfigMaps
:secret Secrets
:sa ServiceAccounts
:pvc PersistentVolumeClaims
:pv PersistentVolumes
:sc StorageClasses

RBAC

Command Resource
:rb RoleBindings
:crb ClusterRoleBindings
:cr ClusterRoles
:role Roles

Auto-scaling

Command Resource
:hpa HorizontalPodAutoscalers
:vpa VerticalPodAutoscalers

Advanced Views

Command Description
:crd All Custom Resource Definitions
:xray deploy default Full dependency tree for deployments in default ns
:pulses Live cluster-wide CPU/MEM metrics dashboard
:popeye Run cluster sanitizer (finds misconfigs, deprecated APIs)

5 Workflows That Will Save You Hours

🔥 1. Debug a CrashLoopBackOff Pod 

1. :pod           → open pod view
2. /api-gateway   → filter by name
3. l              → stream current logs
4. p              → view previous container logs (before crash)
5. d              → describe pod, check Events section
6. s              → shell in if the pod is running
Enter fullscreen mode Exit fullscreen mode

All without leaving K9s, all without remembering a single pod name.

📈 2. Investigate a Memory Spike 

1. Ctrl+P         → open the pulse metrics panel
2. :pod           → go to pod view
3. Shift+M        → sort by memory usage (highest first)
4. Select offender → Enter → select container
5. s              → shell in and run `top` or `ps aux`
Enter fullscreen mode Exit fullscreen mode

🔁 3. Rolling Restart a Deployment 

1. :deploy        → open deployments view
2. Navigate to deployment
3. Ctrl+L         → triggers rolling restart instantly
4. Enter          → drill into pods and watch new ones come up live
Enter fullscreen mode Exit fullscreen mode

No more copy-pasting: kubectl rollout restart deployment/my-app -n production

🌍 4. Test a Service Locally with Port-Forward 

1. :svc           → open services view
2. Select service
3. Shift+F        → port-forward dialog opens
4. Enter local port (e.g., 8080)
5. Access at localhost:8080 in your browser
6. :pf            → manage or stop active port-forwards anytime
Enter fullscreen mode Exit fullscreen mode

⚠️ 5. Drain a Node for Maintenance 

1. :node          → open nodes view
2. Select the node to maintain
3. Ctrl+U         → cordon it (no new pods scheduled)
4. Watch pods reschedule to other nodes in real time
5. Ctrl+D         → drain the node (evict remaining pods)
6. Do your maintenance work
7. Ctrl+N         → uncordon to bring it back
Enter fullscreen mode Exit fullscreen mode

Pro Tips & Power User Tricks

🎨 Custom Skins / Themes

K9s supports fully custom color themes. Place a skin YAML in ~/.config/k9s/skins/.

Popular community themes:

  • Dracula — dark purple theme
  • Tokyo Night — muted dark blues
  • Catppuccin — soft pastel palette
  • Monokai — classic terminal colors

All available in the K9s skins repo.

🔌 Custom Plugins

Create ~/.config/k9s/plugins.yaml to add custom commands to any resource view.

Example — tail logs with stern:

plugin:
  stern:
    shortCut: Shift-S
    description: "Stern multi-pod tail"
    scopes:
      - pods
    command: stern
    background: false
    args:
      - $NAME
      - -n
      - $NAMESPACE
Enter fullscreen mode Exit fullscreen mode

Example — open resource in VS Code:

plugin:
  vscode:
    shortCut: Shift-V
    description: "Open YAML in VSCode"
    scopes:
      - all
    command: bash
    background: false
    args:
      - -c
      - "kubectl get $RESOURCE_NAME $NAME -n $NAMESPACE -o yaml | code -"
Enter fullscreen mode Exit fullscreen mode

🔭 X-Ray View — Dependency Trees

One of K9s's most powerful hidden features:

:xray deploy default
Enter fullscreen mode Exit fullscreen mode

This renders the full dependency tree of all deployments in the default namespace:

Deployment/api-gateway
  └── ReplicaSet/api-gateway-7d9f8b
        ├── Pod/api-gateway-7d9f8b-xkp2q
        │     ├── Container/api-gateway (Running)
        │     └── Container/sidecar (Running)
        └── Pod/api-gateway-7d9f8b-r9pmk
              └── Container/api-gateway (Running)
Enter fullscreen mode Exit fullscreen mode

Perfect for understanding resource relationships and ownership chains.

🧹 Popeye — Cluster Sanitizer 

:popeye
Enter fullscreen mode Exit fullscreen mode

Popeye scans your cluster for:

  • 🔴 Misconfigured resources
  • 🟡 Over-provisioned CPU/memory requests
  • 🟠 Deprecated API versions in use
  • 🔵 Missing resource limits
  • 🔐 Security risks (privileged containers, host mounts)

It gives each resource a grade (A–F) so you know exactly what to fix.

⚙️ K9s Configuration File

K9s config lives at ~/.config/k9s/config.yaml:

k9s:
  refreshRate: 2           # seconds between refreshes
  maxConnRetry: 5
  readOnly: false          # set true to make read-only globally
  noExitOnCtrlC: false
  ui:
    enableMouse: false
    headless: false
    logoless: false
    crumbsless: false
    reactive: false
    noIcons: false
  skipLatestRevCheck: false
  disablePodCounting: false
  shellPod:
    image: busybox
    namespace: default
    limits:
      cpu: 100m
      memory: 100Mi
Enter fullscreen mode Exit fullscreen mode

🔑 Aliases — Short Names for Resources

Add custom aliases in ~/.config/k9s/aliases.yaml:

aliases:
  dp: deployments
  ing: ingresses
  sts: statefulsets
  pp: v1/pods
  crds: apiextensions.k8s.io/v1/customresourcedefinitions
Enter fullscreen mode Exit fullscreen mode

Conclusion

K9s isn't just a nice-to-have — it's the kind of tool that makes you genuinely faster at your job. Whether you're debugging a production incident at 2am or iterating on a local cluster, K9s cuts through the ceremony and gets you to the information you need instantly.

Why K9s belongs in every engineer's toolkit:

  • ⚡ Speed — keyboard shortcuts beat kubectl commands every time
  • 👁️ Visibility — real-time view of your entire cluster in one screen
  • 🔍 Debugging — logs, shell, describe, events — all in two keystrokes
  • 🛡️ Safety — --readonly mode for production, RBAC-aware by default
  • 🧩 Extensible — plugins, skins, aliases, custom shortcuts
  • 🆓 Free & Open Source — Apache 2.0 license, actively maintained

Install it today. The first time you press l to tail logs and then s to shell in — without typing a single pod name — you'll understand why 30,000+ engineers consider K9s essential.

brew install derailed/k9s/k9s && k9s
Enter fullscreen mode Exit fullscreen mode

🔗 Resources

If this helped you, drop a ❤️ and share it with your team. Found a shortcut I missed? Drop it in the comments below! 👇

Happy clustering! 🐕

Top comments (0)