Skip to main content

Scenario Lab

Interactive diagnosis demos grounded in real Kubernetes failure fixtures.

Edit the manifest locally, inspect the cluster signals, and see how triage presents the incident. The sandbox stays deterministic by design: outputs remain curated to the selected scenario.

Manifest input

A pod restarts fast enough that the incident starts as noise and becomes an outage.

kubediag promotes the runtime failure to the top, preserves the raw evidence, and gives the exact commands needed to confirm the crash cause.

This panel starts from a repository-backed fixture and can be edited safely in-browser.

Cluster signals

What triage would anchor on first

  • pod.status.containerStatuses[app].state.waiting.reason = CrashLoopBackOff
  • pod.status.containerStatuses[app].restartCount = 5
  • Warning BackOff: back-off restarting failed container
  • Last 3 log lines show the process exits almost immediately

Diagnosis output

TRG-POD-CRASHLOOPBACKOFF

triage-output:text
▶ Pod default/crashloop-demo
  generated: 10:30:00 UTC  (42ms)
  overall: CRITICAL
  findings: 2 (1 critical, 1 high)

ⓧ CRITICAL  [high confidence]  TRG-POD-CRASHLOOPBACKOFF
   Container "app" is in CrashLoopBackOff (5 restarts)
   Container "app" has crashed 5 times. The kubelet is backing off restarts.

   Evidence:
     • pod.status.containerStatuses[app].state.waiting.reason = CrashLoopBackOff
     • pod.status.containerStatuses[app].restartCount = 5

   Next commands:
     $ kubectl logs -n default crashloop-demo -c app --previous
     $ kubectl describe pod -n default crashloop-demo

● HIGH      [high confidence]  TRG-POD-OOMKILLED
   Container "app" was OOMKilled
   Container "app" was killed by the kernel OOM killer.

Why this finding wins

  • The waiting reason is explicit and high-confidence; kubediag can name the runtime symptom without guessing.
  • Restart count confirms this is not a one-off crash but a repeating failure pattern.
  • Related findings, like OOMKilled, stay visible but ranked below the primary incident driver.

Next commands

Pasteable follow-up for responders

  • kubectl logs -n default crashloop-demo -c app --previous
  • kubectl describe pod -n default crashloop-demo