Model Context Protocol

About the Model Context Protocol integration

@telemetry-dev/mcp instruments transports from the official TypeScript MCP SDK v2. It records client and server request spans, correlates JSON-RPC responses, and carries W3C trace context through MCP params._meta, including across stdio process boundaries.

Key features

• End-to-end MCP traces: Client and server spans share one trace even when MCP runs over stdio.
• Protocol details: Inspect methods, request IDs, protocol versions, sessions, tools, prompts, resources, latency, and errors.
• Tool-aware spans: tools/call requests appear as standard execute_tool spans with the tool name.
• Correct error classification: JSON-RPC errors, tool failures, transport failures, and connection closure are recorded without treating caller mistakes as server faults.
• Private by default: Tool arguments, results, and OpenTelemetry baggage stay local unless you enable their respective capture or propagation options.

Get started

Set TELEMETRY_DEV_API_KEY, initialize telemetry.dev, then instrument the transport before connecting it:

import { Client } from "@modelcontextprotocol/client";
import { StdioClientTransport } from "@modelcontextprotocol/client/stdio";
import { instrumentMcpTransport } from "@telemetry-dev/mcp";
import { init } from "@telemetry-dev/sdk";

init({ serviceName: "mcp-client" });

const client = new Client({ name: "research-agent", version: "1.0.0" });
const transport = instrumentMcpTransport(
  new StdioClientTransport({ command: "research-mcp-server" }),
);

await client.connect(transport);

Use the same instrumentMcpTransport() call with a server transport before server.connect(). Pass { capturePayloads: true } only when tool arguments and successful results may be collected under your project's privacy policy.

The package supports MCP SDK v2. MCP v1 and MCP-specific metrics are not currently instrumented.