| name | container-crash-analysis |
| description | Analyzes crashing, restarting, or OOMKilled Docker/Kubernetes containers. Fixes crash causes. Do NOT use for deployment failures, app bugs, or infrastructure.
|
| version | 1.1.0 |
| time_saved | Manual: 1-3h | With skill: 15-30m |
| license | Proprietary — Personal Use Only |
| category | debugging |
| complexity | Advanced |
| tokens | ~5000 |
| tags | ["container","docker","kubernetes","pod","oomkilled","crash","debugging"] |
| author | vheins |
Skill: Container Crash Analysis
Purpose
Analyze container/pod failures to identify root causes (OOM, probes, config) and provide targeted fixes.
Input
| Variable | Type | Req | Description |
|---|
tech_stack | string | Yes | e.g., "Docker + Kubernetes" |
container_logs | string | Yes | Logs, symptoms, or exit codes |
pod_description | string | No | kubectl describe pod output |
context | string | Yes | Limits, changes, crash frequency |
Instructions
- Classification: Identify type (OOMKilled/137, App Crash/1, Startup, Liveness, CrashLoop, Segfault/139).
- Analysis: Pinpoint memory-consuming process, exception trace, or missing config.
- Remediation: Provide corrected manifests/configs; adjust probe timings.
- Tuning: Right-size resources (Requests = 50-70% of limits) using
top pod or VPA.
- Monitoring: Setup alerts for restarts (>3/10m), OOM, and CrashLoop states.
- Fallback: If no logs, provide retrieval commands and crash-type checklist.
Edge Cases
| Case | Strategy |
|---|
| Intermittent | Recommend persistent monitoring for leaks; check graceful shutdown. |
| Init Containers | Use kubectl logs -c for init; diagnose as a separate block. |
| High Load OOM | Recommend Vertical Pod Autoscaler (VPA) or horizontal scaling. |
Diagnostic Workflow
flowchart TD
A([Start: Container Crash Analysis]) --> B[Parse inputs]
B --> C{Logs available?}
C -- No --> D[Provide retrieval commands]
D --> E[Describe crash types]
C -- Yes --> F[Identify Crash Type]
F --> G{Type?}
G -- OOMKilled 137 --> H[Identify memory-consuming process]
G -- App crash 1 --> I[Find exception/panic + stack trace]
G -- Startup failure --> J[Check config/port/dependency]
G -- Liveness failure --> K[Check endpoint/port/startup]
G -- CrashLoopBackOff --> L[Identify repeating crash]
G -- Segfault 139 --> M[Analyze core dump]
H --> N[Increase limit/fix leak]
I --> O[Fix exception/handling]
J --> P[Add config/port/init container]
K --> Q[Adjust timing/endpoint/startupProbe]
L --> R[Fix root cause]
M --> S[Check native deps]
N & O & P & Q & R & S --> T[Show corrected config]
T --> U{Resource crash?}
U -- Yes --> V[Right-size:\nkubectl top\nRequests=50-70%\nVPA]
U -- No --> W
V --> W[Setup alerts:
Restarts>3/10m
OOMKilled
CrashLoopBackOff]
W --> X([Output: 5 Sections + Config + Monitoring])
Examples
Quality Gate
MCP Dependencies
@upstash/context7-mcp: Library documentation and examples.
Changelog
| Version | Date | Description |
|---|
| 1.1.0 | 2026-03-20 | Restructured: moved examples, references, added compatibility/license |
| 1.0.0 | 2026-03-20 | Initial release |
