Command Speed Tactics Under Pressure

Why this matters

Seven minutes per task leaves no budget for hand-written YAML or from-scratch recall. Every second spent typing manifests or remembering field names is stolen from diagnosis, and diagnosis is where the points live. The toolkit from M00 plus the loops from every module compose into a pace that feels unhurried because nothing is invented mid-task.

Concepts

Generate, never author: kubectl create ... -o yaml --dry-run=client plus kubectl run variants produce valid manifests in seconds, edited only where the task differs. Docs are extended memory: bookmark the task pages (deployments, services, RBAC, storage) and search field names there instead of recalling them. Explain closes the loop: kubectl explain answers nesting questions faster than any search.

Aliases plus completion load first: alias k=kubectl, shell completion, and three template lines (create-with-dry-run, expose, scale) pasted from notes in minute zero. Output discipline saves more: -o wide, -o yaml for exact fields, jsonpath for single values, --sort-by for events. Each saves a re-run.

Worked example

A service-debugging minute, toolkit pace:

k get svc,deploy,pods -o wide          # one call, whole picture
k get endpoints api                     # plumbing verdict in seconds
k explain svc.spec.ports | head         # field doubt resolved, no search
k exec deploy/api -- nslookup db        # inside view, one line
k get events --sort-by=.lastTimestamp | tail -5   # platform says its piece
# total: under 90 seconds from symptom to layer, diagnosis next

No manifest typed, no field recalled, no tab-hopping. The diagnosis that follows spends the saved minutes where they pay.

Exam notes

Templates must be tested before exam day: an untested alias fails exactly when needed. Docs searching needs practice too: finding a field in under a minute is trained in the M00 drill, not discovered mid-exam. And wide output first, always: the extra columns prevent the second command more often than not.