k8s-ai-cli 0.1.1

CLI that explains Kubernetes pod failures from logs using AI. Works locally with kubectl.

k8s-ai

CLI that explains Kubernetes pod failures from logs using AI.

Features

• Collects logs from Kubernetes pods via kubectl
• AI-powered root-cause analysis (OpenAI-compatible API)
• Detects common failure classes:
  • CrashLoopBackOff
  • OOMKilled
  • ImagePullBackOff
  • DB connection timeout/errors
  • Permission denied / RBAC-style issues
  • HTTP 500 style failures
• Suggests remediation steps and safe kubectl follow-up commands
• Supports multi-container pods and --previous logs
• Offline --mock mode (no kubectl calls required)
• Sensitive data redaction before model input

Project Structure

k8s_ai/
├── .env.example
├── pyproject.toml
├── requirements.txt
├── README.md
└── k8s_ai/
    ├── __init__.py
    ├── __main__.py
    ├── cli.py
    ├── config.py
    ├── models.py
    ├── logging_setup.py
    ├── collector/
    ├── analyzer/
    ├── fixer/
    └── utils/

Setup

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env

Update .env values:

• LLM_API_KEY
• LLM_BASE_URL
• LLM_MODEL

Usage

Analyze pod:

python -m k8s_ai analyze <namespace> <pod>

Analyze specific container:

python -m k8s_ai analyze <namespace> <pod> --container <container-name>

Use previous container logs:

python -m k8s_ai analyze <namespace> <pod> --previous

Run fully offline:

python -m k8s_ai analyze default dummy-pod --mock

Notes

• In non-mock mode, kubectl and cluster access are required.
• In mock mode, realistic sample logs are used and no kubectl command is executed.

License

MIT

Kubernetes Log Analyzer Agent

AI-powered CLI tool to collect Kubernetes pod logs and perform root-cause analysis with safe, structured remediation guidance.

Features

• Analyze logs for any namespace/pod
• Multi-container pod support
• Previous container logs (--previous)
• Timeout handling
• Intelligent log truncation
• Secret redaction before LLM call
• Optional suggested kubectl fix commands
• Offline mock mode for testing without API calls
• Rich terminal output

Project Structure

k8s_ai/
├── .env.example
├── requirements.txt
├── README.md
└── k8s_ai/
    ├── __init__.py
    ├── __main__.py
    ├── cli.py
    ├── config.py
    ├── models.py
    ├── logging_setup.py
    ├── collector/
    │   ├── __init__.py
    │   └── kubectl_collector.py
    ├── analyzer/
    │   ├── __init__.py
    │   └── llm_analyzer.py
    ├── fixer/
    │   ├── __init__.py
    │   └── command_generator.py
    └── utils/
        ├── __init__.py
        ├── redaction.py
        └── truncation.py