K9s Cheat Sheet: Essential Commands and Troubleshooting Guide

K9s is a terminal based UI to interact with your Kubernetes clusters. K9s continually watches Kubernetes for changes and offers subsequent commands to interact with your observed resources.

Installation

Linux

wget https://github.com/derailed/k9s/releases/download/v0.32.7/k9s_linux_amd64.deb && apt install ./k9s_linux_amd64.deb && rm k9s_linux_amd64.deb

macOS

brew install derailed/k9s/k9s

Navigation Basics

Before diving into troubleshooting, it’s important to understand k9s navigation:

Essential Commands

• Help Menu: ? - Access the help documentation
• Exit K9s: :q or Ctrl+C - Quit the application
• Resource Aliases: Ctrl+A - View available resource shortcuts
• Filtering:
  • Basic Filter: / followed by your search term
  • Inverse Filter: /! followed by exclusion term
  • Label Filter: /-l followed by label selector

Working with Kubernetes Controllers

StatefulSets (sts)

• View & Filter:
  • View: Use :sts command
  • Filter by name: :sts /name
  • View in specific namespace: :sts namespace-name
• Common operations:
  • Scale: Select StatefulSet and press s
  • Describe: Press d
  • Edit: Press e
  • Delete: Press ctrl-d (with confirmation)

DaemonSets (ds)

• View & Filter:
  • View DaemonSets: Use :ds command
  • Filter DaemonSets: :ds /name
  • View in specific namespace: :ds namespace-name
• Common operations:
  • View pods: Select DaemonSet and press enter
  • Describe: Press d
  • Edit: Press e
  • Delete: Press ctrl-d (with confirmation)

ReplicaSets (rs)

• View & Filter:
  • View ReplicaSets: Use :rs command
  • Filter ReplicaSets: :rs /name
  • View in specific namespace: :rs namespace-name
• Common operations:
  • Scale: Select ReplicaSet and press s
  • Describe: Press d
  • Edit: Press e
  • Delete: Press ctrl-d (with confirmation)

Troubleshooting Missing Pods

When a pod doesn’t appear in the expected k9s view:

1. Check other resource types:

   • Use :sts, :ds, or :rs commands
   • View specific resources using singular/plural forms: :pod or :pods
   • Filter pods by name: :pod /name
   • Filter by labels: :pod app=name,env=dev

2. View all resources:

   • Use :all command
   • Use :screendump or :sd to view all saved resources

3. Verify the correct namespace:

   • Switch namespaces: Use :ns
   • View pods in specific namespace: :pod namespace-name

4. Examine pod status:

   • Use :pods to find and check pod states
   • Use d to describe resources
   • Use l to view logs
   • Use v to view resource details
   • Use e to edit resources

If found, but not as a Deployment, the pod might be:

• Orphaned from a deleted Deployment
• Managed directly without a controller

Scaling in k9s

To manually scale (once found in Deployments view):

1. Highlight the deployment
2. Press s
3. Enter desired replica count
4. Press Enter to confirm

Important Notes on Scaling

• Manual scaling in k9s doesn’t update Helm charts or releases
• Scaled values persist until the next deployment or manual change
• Helm deployments may override manual scaling if replica count is specified in the chart
• To preserve manual scaling across Helm upgrades, use --reuse-values flag (use cautiously)
• For permanent changes, update replica count in Helm chart values

Advanced Operations

1. Context Management:

   • Switch contexts: Use :ctx
   • Direct context switch: :ctx context-name
   • View pods in specific context: :pod @ctx1

2. Resource Management:

   • Delete resource: Press ctrl-d (requires confirmation)
   • Force delete resource: Press ctrl-k (no confirmation)
   • View resource relationships: Use :xray RESOURCE [NAMESPACE]
   • Monitor resource metrics: Use :pulses or :pu

Useful Links

• k9s GitHub Repository
• Kubernetes Deployments Documentation
• Helm Documentation
• Kubernetes Scaling Documentation

Previous

← Kubernetes Workload Types

Next

The Case of the Missing Makefile: A Git Story →