AI Log Analyzer

A powerful log analysis tool that combines system log collection with AI-powered insights. Supports local and remote log collection, Docker container logs, and multiple AI providers (Claude, ChatGPT, and Gemini).

Features

• Log Collection

  • Local system logs
  • Remote system logs via SSH
  • Docker container logs
  • Support for custom Docker socket paths

• Flexible Filtering

  • Filter by severity (error, warning, info)
  • Group similar messages
  • Exclude specific message types
  • Summary view option

• AI Analysis

  • Multiple AI provider support (Claude, ChatGPT, Gemini)
  • Comparative analysis across providers
  • Customizable analysis depth
  • Save analysis results to file

• Docker Integration

  • Container log collection
  • Container stats collection
  • Custom socket path support
  • Container filtering

Installation

1. Clone the repository:

git clone https://github.com/alt229/ai_log_analyzer.git
cd ai-log-analyzer

2. Install dependencies:

pip install -r requirements.txt

Configuration

Before using the AI features, you'll need to set up your API keys:

# Set API keys
python main.py --set-api-key claude your-claude-key
python main.py --set-api-key chatgpt your-openai-key
python main.py --set-api-key gemini your-google-key

# View current configuration
python main.py --show-config

# Reset to defaults
python main.py --reset-config

Configuration is stored in ~/.config/ai_logs/config.yaml.

Usage Examples

Basic Log Analysis

# Analyze local logs for the last hour
python main.py -t 1

# Show only errors
python main.py -t 1 --only-errors

# Full message display
python main.py -t 1 --full

# Summary view
python main.py -t 1 --summary

Remote Log Analysis

# Analyze remote host logs
python main.py -t 1 --host server1.example.com --user admin --key ~/.ssh/id_rsa

# With AI analysis
python main.py -t 1 --host server1.example.com --user admin --key ~/.ssh/id_rsa --ai claude

Docker Log Analysis

# Analyze all container logs
python main.py -t 1 --host server1.example.com --user admin --docker

# Specific container
python main.py -t 1 --host server1.example.com --user admin --docker --container nginx

# Custom Docker socket
python main.py -t 1 --host server1.example.com --user admin --docker --docker-socket /run/user/1000/docker.sock

AI Analysis

# Basic AI analysis
python main.py -t 1 --ai claude

# Compare multiple AI providers
python main.py -t 1 --ai claude --compare

# Save analysis to file
python main.py -t 1 --ai claude -o analysis.txt

Options

Time options:
  -t, --time HOURS         Number of hours to look back (default: 1)

Remote connection options:
  --host HOST             Remote host to analyze logs from
  --user USER             SSH username for remote host
  --port PORT             SSH port (default: 22)
  --key KEY              Path to SSH private key file

Output options:
  --no-color             Disable colored output
  --json                 Output results as JSON
  --full                 Show full messages without truncation
  -o, --output FILE      Write output to file
  --debug                Show debug information
  --summary              Show summarized output

Docker options:
  --docker               Include Docker container logs
  --container NAME       Specific container to analyze
  --no-container-stats   Skip container stats collection
  --docker-socket PATH   Path to Docker socket

AI analysis options:
  --ai PROVIDER          Enable AI analysis (claude, gemini, or chatgpt)
  --api-key KEY          API key for chosen AI provider
  --system-info FILE     Path to JSON file containing system information
  --compare              Compare analyses from all configured AI providers

Requirements

• Python 3.8+
• Required packages (see requirements.txt):
  • pyyaml
  • colorama
  • paramiko
  • anthropic
  • openai
  • google-generativeai
  • docker

pip install -r requirements.txt

Contributing

Contributions are welcome! Please feel free to submit pull requests.

ToDo

Add windows event viewer support

Authors

Erin Scott (but mostly Claude 3.5)

Acknowledgments

• Thanks to Anthropic, OpenAI, and Google for their AI APIs!