kubernetes-log-analysis 0.1.0

A tool to analyze Kubernetes log bundles using LLMs.

Kubernetes Log Analysis Tool (k-log)

中文版說明文件

A command-line tool that utilizes Large Language Models (LLMs) to analyze collected Kubernetes log content.

Prerequisites

• Python 3.8+
• A directory of Kubernetes log content (generated by k8s-debug.sh, for details see https://github.com/ap-mic-inc/kubernetes-log-analysis).
• Google Gemini API Key (or an API key for other models supported by LiteLLM).

Python Package Installation

1. Install the package and its dependencies using pip:

   pip install kubernetes-log-analysis==0.1.0
   

Configuration

1. API Key: This tool requires the GOOGLE_API_KEY for Gemini. Create a file named .env in the directory where you will run k-log (or the project root during development) with the following content:

   GOOGLE_API_KEY="YOUR_GEMINI_API_KEY_HERE"
   

   Alternatively, ensure this environment variable is set globally.

2. LLM Model (Optional): You can specify a different LiteLLM compatible model string by setting the LLM_MODEL environment variable in your .env file or globally. The default is gemini/gemini-2.5-flash-preview-04-17.

   LLM_MODEL="gemini/gemini-2.5-flash-preview-04-17"
   

Usage

1. Ensure your GOOGLE_API_KEY is set (e.g., via the .env file).

2. Navigate to your Kubernetes log bundle directory (e.g., cd /path/to/k8s-debug-default-20250509034522).

3. Run the k-log command with your query:

   k-log "What is the current status of the nodes?"
   

   Alternatively, explicitly specify the log directory:

   k-log --log-dir /path/to/your/k8s-debug-default-20250509034522 "logs for pod kuberay-operator-6ddb59999c-mggc9"
   

Example Queries:

• k-log "status of nodes"
• k-log "logs for pod jupyterlab-54698b59b8-bmvgq in namespace default"
• k-log "describe service kubernetes"
• k-log "any critical events?"

About the Author

Simon Liu

APMIC MLOps Engineer x Google Developer Expert (GDE) in AI

A technology enthusiast in the field of artificial intelligence solutions, focusing on assisting enterprises in adopting generative AI, MLOps, and large language model (LLM) technologies to drive digital transformation and practical technological implementation.

Currently also a Google Developer Expert (GDE) in the GenAI field, actively participating in technology communities. Through technical articles, speeches, and practical experience sharing, he promotes the application and development of AI technology. To date, he has published over a hundred technical articles on Medium, covering topics such as generative AI, RAG, and AI Agents, and has served multiple times as a speaker at technical seminars, sharing practical applications of AI and generative AI.

Related Links:

• APMIC Official Website: https://www.apmic.ai/
• Personal Social Media Links: https://simonliuyuwei.my.canva.site/link-in-bio