Model Context Protocol monitoring
Gain visibility into Model Context Protocol (MCP) servers and clients with distributed tracing for tool execution, request flows, and performance of MCP applications.
Overview
Instana provides comprehensive observability for MCP servers and clients. MCP is a standard that enables AI assistants like Claude Desktop and GitHub Copilot to securely connect to external data sources and tools.
With Instana's MCP monitoring, you gain visibility into tool execution, request flows, and performance of your MCP applications. The integration uses OpenLLMetry to provide distributed tracing for MCP operations, automatically generating spans for key activities that include tool execution, request processing, and client/server interactions.
Key features
MCP monitoring includes the following key features:
- Distributed tracing: Track MCP requests from client through server to tool execution.
- Tool execution monitoring: Monitor performance and success rates of MCP tools.
- Request and response tracking: View complete request and response payloads.
- Error detection: Automatically capture and analyze MCP errors.
- Performance metrics: Track latency, throughput, and resource usage.
- Client/server correlation: Understand interactions between MCP clients and servers.
Common use cases
See the following example use cases:
- MCP server performance
- Monitor how your MCP servers handle requests from AI assistants. Identify slow tools, optimize request handling, and make sure that the interactions are responsive.
- Tool usage analysis
- Understand which MCP tools are used most frequently, which are slow, and which fail most often. Optimize tool implementations based on actual usage patterns.
- Error troubleshooting
- When MCP tools fail or return errors, view complete traces that show the request flow, parameters, and error details. Quickly identify and fix issues.
- Integration monitoring
- Monitor MCP integrations with AI assistants. Make sure that the connections are reliable and detect integration issues early.
Where to find it in the UI
You can access MCP monitoring from the following locations in the UI:
- Gen AI observability dashboard: Click Gen AI Observability in the navigation menu to access MCP traces and performance data.
- Traces view: Click Gen AI Observability and select the Traces tab to view individual MCP operations with request details, tool execution, and errors.
- Analytics view: Click Analyze gen AI calls to analyze MCP calls by application, service, and endpoint.
What problems does this solve?
MCP monitoring addresses the following common challenges with MCP-based applications:
- MCP server analysis: Gain visibility into what happens inside MCP servers.
- Slow tools: Identify and optimize slow tool executions.
- Error diagnosis: Understand why MCP tools fail with complete context.
- Performance issues: Detect performance problems that affect AI assistant responsiveness.
- Usage insights: Understand how AI assistants use your MCP tools.
Related capabilities
The following capabilities work together with MCP monitoring:
- Distributed tracing - Core tracing technology for MCP monitoring
- Generative AI observability - Monitor AI applications by using MCP
- Application perspectives - Organize MCP services
- Smart Alerts - Alert on MCP performance issues
Learn more
For detailed information about MCP monitoring, instrumentation, and configuration, see Monitoring MCP servers.