
Performance Co-Pilot
Connects to Performance Co-Pilot's pmproxy REST API to expose system metrics through five MCP tools: get_system_snapshot for CPU, memory, disk, and network overview; get_process_top for sorting processes by resource usage; query_metrics for fetching specific PCP metrics; search_metrics for discovery; and describe_metric for metadata. Built on FastMCP, runs via uvx without installation, and works against localhost pmcd by default or any remote host via PCP_TARGET_HOST. You'd reach for this when troubleshooting performance issues or capacity planning, asking Claude to analyze bottlenecks, compare resource trends, or identify which processes are consuming resources. Requires PCP with pmcd and pmproxy running on the monitored systems.
- tools
- 5
- transport
- STDIO
- env vars
- 3
- last updated
- Mar 26, 2026
Tools
5 toolsPublic tool metadata for what this MCP can expose to an agent.
get_system_snapshotPoint-in-time system overview (CPU, memory, disk, network, load)
Point-in-time system overview (CPU, memory, disk, network, load)
No parameter schema in public metadata yet.
get_process_topTop processes by CPU, memory, or I/O usage
Top processes by CPU, memory, or I/O usage
No parameter schema in public metadata yet.
query_metricsFetch current values for specific PCP metrics
Fetch current values for specific PCP metrics
No parameter schema in public metadata yet.
search_metricsDiscover available metrics by name pattern
Discover available metrics by name pattern
No parameter schema in public metadata yet.
describe_metricGet detailed metadata about a metric
Get detailed metadata about a metric
No parameter schema in public metadata yet.
Sponsors
Configuration
PCP_HOSTpmproxy host
PCP_PORTpmproxy port
PCP_TARGET_HOSTTarget pmcd host to monitor
Readme
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.
π Full Documentation | π Getting Started
π Quick Start (No Install)
Run immediately with uvx β no installation required:
uvx pcp-mcp
Or install as a persistent global tool:
uvx tool install pcp-mcp
pcp-mcp
π¦ Installation
pip install pcp-mcp
Details
- Repository
- github.com/major/pcp-mcp
- Category
- AI & LLM Tools
- Package
pcp-mcp- Transport
- STDIO
- Registry
- MCP registry, active