Kubernetes Monitor CLI (kai)

Overview

Kubernetes Monitor CLI (kai) is a powerful command-line tool that helps you diagnose and troubleshoot Kubernetes clusters:

• Monitor pod health across multiple namespaces
• Diagnose pod issues with AI-powered analysis
• Get actionable recommendations for resolving common Kubernetes problems
• View relevant logs and events in a structured format

Features

• Command-line interface: Simple and intuitive CLI for monitoring Kubernetes resources
• Multi-namespace monitoring: Monitor pods across one or more namespaces
• AI-powered analysis: Uses Claude 3.7 Sonnet to analyze issues and provide human-readable diagnoses
• Rich terminal output: Structured and color-coded display of issues, logs, and events
• Custom kubeconfig support: Specify custom kubeconfig paths for different environments

Installation

Prerequisites

• Python 3.9+
• Access to Kubernetes clusters
• Anthropic API key (optional, for AI-powered analysis)

Setup

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt
   

3. (Optional) Set environment variables for AI analysis:

   export ANTHROPIC_API_KEY="your-api-key"
   

Usage

Run the CLI tool directly:

python main.py [command] [options]

Or create an alias for easier access:

alias kai="python /path/to/main.py"

Or compile the application:

pyinstaller main.spec
ln -s dist/main/main /usr/local/bin/kai

Commands

info

Display information about the tool and available commands.

kai info

pods

Monitor pod health in specified namespaces.

kai pods [--kubeconfig PATH] [--namespace NAMESPACE1,NAMESPACE2]

Options:

• --kubeconfig, -k: Path to the kubeconfig file (defaults to ~/.kube/config)
• --namespace, -n: Kubernetes namespaces to monitor (defaults to default)

Monitored Kubernetes Issues

Kubernetes Monitor CLI tracks the following types of pod issues:

Pod Status Issues

• CrashLoopBackOff: Pods that continuously crash and restart
• ImagePullBackOff: Pods that fail to pull their container images
• CreateContainerError: Pods that fail during container creation
• OOMKilled: Pods terminated due to out-of-memory conditions
• Terminated with Error: Containers that exit with non-zero status codes
• Pending/Failed/Unknown: Pods stuck in problematic phases

Resource Issues

• Container Restarts: Containers that have excessive restart counts
• Resource Constraints: Pods hitting CPU/memory limits

AI Analysis

When an Anthropic API key is provided, each issue is analyzed by Claude 3.7 Sonnet to provide:

• Concise diagnosis of the root cause
• Specific, actionable recommendations to resolve the issue
• Context-aware analysis based on pod logs and events