pcp-mcp¶

MCP server for Performance Co-Pilot (PCP) metrics.

Query system performance metrics via the Model Context Protocol - CPU, memory, disk I/O, network, processes, and more.

Features¶

System Monitoring - CPU, memory, disk I/O, network, load averages
Process Monitoring - Top processes by CPU, memory, or I/O usage
Metric Discovery - Search and explore 1000+ PCP metrics
Remote Monitoring - Monitor any host running pmcd
Real-time Data - Direct access to PCP's high-resolution metrics
MCP Resources - Browse metrics via pcp:// URIs

Architecture¶

┌─────────┐         ┌─────────┐          ┌─────────┐         ┌─────────┐
│   LLM   │ ◄─MCP─► │ pcp-mcp │ ◄─HTTP─► │ pmproxy │ ◄─────► │  pmcd   │
└─────────┘         └─────────┘          └─────────┘         └─────────┘
                                         (REST API)          (metrics)

pcp-mcp is a FastMCP server that exposes Performance Co-Pilot metrics through the Model Context Protocol. It connects to PCP's REST API server (pmproxy) to fetch real-time system metrics.

Quick Start¶

# Install
pip install pcp-mcp

# Install PCP (if not already installed)
sudo dnf install pcp
sudo systemctl enable --now pmproxy

# Run the MCP server
pcp-mcp

Configure Claude Desktop (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "pcp": {
      "command": "pcp-mcp"
    }
  }
}

Use Cases¶

🔍 Performance Troubleshooting¶

Ask your LLM: - "Analyze current system performance and identify bottlenecks" - "Why is my disk I/O so high?" - "Which processes are consuming the most CPU?"

📊 System Monitoring¶

"Give me a health check of the production server"
"Compare CPU usage over the last minute"
"Monitor network traffic on eth0"

📈 Capacity Planning¶

"What's the memory utilization trend?"
"Show me disk usage across all filesystems"
"Analyze process resource consumption patterns"

Requirements¶

Python: 3.10+
PCP: Performance Co-Pilot with pmproxy running