Pulse/docs/AI.md

Pulse AI

Pulse AI is an intelligent monitoring assistant that helps you understand and manage your infrastructure through natural language conversations, proactive monitoring, and automated issue resolution.

Features

Chat Assistant

Ask questions about your infrastructure in plain English:

• "What's using the most CPU right now?"
• "Show me VMs with high memory usage"
• "Why is my backup failing?"
• "Summarize the health of my cluster"

Patrol Mode

Automated monitoring that runs on a schedule to:

• Analyze resource utilization patterns
• Detect potential issues before they cause problems
• Generate actionable recommendations
• Track trends over time

Auto-Fix

Automatically resolve common issues:

• Restart stuck services
• Clear disk space
• Restart unresponsive containers
• Apply recommended optimizations

Predictive Intelligence

• Identify resources trending toward problems
• Forecast disk space exhaustion
• Detect unusual patterns
• Alert on anomalies

Configuration

Enable Pulse AI

1. Navigate to Settings → AI
2. Toggle Enable Pulse AI
3. Configure your AI provider

Supported Providers

Provider	Models	Notes
OpenAI	GPT-4, GPT-4 Turbo, GPT-3.5	Recommended for best results
Anthropic	Claude 3 Opus, Sonnet, Haiku	Excellent for complex analysis
Ollama	Llama 3, Mistral, etc.	Self-hosted, privacy-focused
OpenRouter	Multiple models	Pay-per-use routing
Custom	Any OpenAI-compatible API	For enterprise deployments

API Key Setup

# Environment variable (recommended for production)
export PULSE_AI_PROVIDER=openai
export PULSE_AI_API_KEY=sk-...

# Or configure via Settings UI

Patrol Configuration

Setting	Default	Description
Patrol Enabled	Off	Run automated checks
Patrol Interval	30 minutes	How often to run patrol
Auto-Fix Enabled	Off	Allow automatic remediation
Auto-Fix Model	Same as chat	Model for auto-fix decisions

Usage

Chat Interface

Access the AI chat from the bottom-right corner of any page. Type your question and press Enter.

Example queries:

"What VMs are using more than 80% memory?"
"Show me the status of all backups"
"Why is node pve1 showing high load?"
"Compare resource usage between this week and last week"

Context Selection

Click on any resource (VM, container, host) to add it to the AI context. This helps the AI provide more specific answers.

Patrol Reports

When Patrol is enabled, Pulse AI will:

1. Run periodic health checks
2. Generate findings (issues, warnings, info)
3. Offer to auto-fix issues (if enabled)
4. Track trends over time

Cost Tracking

Track AI API usage in Settings → AI → Cost Dashboard:

• Daily/monthly usage statistics
• Cost breakdown by feature
• Usage trends over time

Privacy & Security

• Data stays local: Only resource metadata is sent to AI providers
• No training: Your data is never used for model training
• Audit logging: All AI interactions are logged
• Self-hosted option: Use Ollama for complete data privacy

Troubleshooting

AI not responding

1. Check API key is configured correctly
2. Verify network connectivity to AI provider
3. Check Settings → AI for error messages

Patrol not running

1. Ensure Patrol is enabled in Settings
2. Check system resource availability
3. Review logs: journalctl -u pulse -f

Auto-fix not working

1. Enable Auto-Fix in Settings → AI
2. Verify the connected agents have execute permissions
3. Check the Auto-Fix model is configured