Kubectl Cheat Sheet

Core Philosophy

1. Optimize for the command you need right now — Cheat sheets should get you to the right command quickly, not explain Kubernetes from first principles.
2. Group by operational task, not alphabetically — Operators think in “inspect pods” and “roll out a deployment,” not in raw verb order.
3. Prefer safe read commands first — get, describe, logs, and diff should come before write actions when debugging.
4. Make selectors and namespaces explicit — Most kubectl mistakes come from targeting the wrong objects or cluster scope.
5. Include output shaping — -o wide, -o yaml, jsonpath, and label selectors often matter as much as the base command.

Quick-Start: Common Problems

"What pods are broken right now?"

1. kubectl get pods -A
2. kubectl describe pod <pod> -n <ns>
3. kubectl logs <pod> -n <ns> --previous -> See references/commands.md

"How do I restart or roll out a deployment?"

Need	Command
restart deployment	kubectl rollout restart deploy/<name> -n <ns>
watch rollout status	kubectl rollout status deploy/<name> -n <ns>
scale replicas	kubectl scale deploy/<name> --replicas=3 -n <ns>
-> See references/commands.md

Decision Trees

Signal	Command family
need to inspect objects	get, describe, logs, top
need to enter or reach workload	exec, port-forward
need to change desired state	apply, delete, scale, rollout
need filtered output	selectors, -o, jsonpath

Reference Index

File	Contents
references/commands.md	kubectl commands grouped by inspect, debug, mutate, rollout, output shaping, namespaces, resources, and common operator tasks