CASE 01
The restarting report worker
Distinguish a restart symptom from a confirmed termination reason.
GUIDED PRACTICE / GETTING STARTED
THE ON-CALL PRACTICE ROOM
Build the habit
of evidence.
Practice the first few minutes of a Kubernetes incident. Inspect the signals, cite your evidence, and explain why your diagnosis fits.
Simulated incidentsNo setupAt your pace
YOUR PRACTICE LOG
— / 10
base cases with submitted checks passed
Keep the evidence. Revisit your reasoning. Try a transfer case.
THE PRACTICE LIBRARY
Small incidents. Better instincts.
10 base cases + 10 transfer cases
CASE 02
Healthy pods, unreachable service
Trace how a Service selects its backend pods.
CASE 03
A release that never starts
Separate registry authentication failures from missing image tags.
CASE 04
A worker exits before serving
Use previous logs to investigate a crash loop.
CASE 05
A Pending pod on a quiet node
Scheduling uses requested capacity, not current CPU usage.
CASE 06
Running but never Ready
Readiness failure can exclude a healthy process from traffic.
CASE 07
Only one client cannot resolve Services
Check the client resolver before blaming CoreDNS.
CASE 08
Configuration prevents container creation
Differentiate startup configuration failure from a crash loop.
CASE 09
A database waits for storage
Use claim events to identify provisioning prerequisites.
CASE 10
The old version stays available
Trace Deployment progress down to ReplicaSet events.
01Read the incident
Separate what you know from what you suspect.
02Collect evidence
Inspect simulated diagnostic results at your own pace.
03Explain the cause
Check your diagnosis and keep the lesson in your notebook.