Add OpenTelemetry instrumentation for MCP (Model Context Protocol)

This commit adds automatic instrumentation for the Model Context Protocol,
enabling distributed tracing for MCP client and server operations.

Key Features:
- Automatic tracing for MCP client requests and notifications
- Automatic tracing for MCP server request/notification handlers
- Distributed trace context propagation via W3C Trace Context
- Support for both stdio and HTTP transports
- MCP semantic conventions support
- Session ID tracking for HTTP transport

Implementation:
- Instruments BaseSession.send_request/send_notification (client-side)
- Instruments Server._handle_request/_handle_notification (server-side)
- Injects/extracts trace context via params._meta field
- Creates linked spans for distributed tracing

Code Quality:
- 58 comprehensive unit tests (100% pass rate)
- Lint score: 10.00/10
- Passes all pre-commit hooks
- Full type annotations and documentation

Package: opentelemetry-instrumentation-mcp
Location: instrumentation-genai/opentelemetry-instrumentation-mcp/
Dependencies: mcp >= 1.8.1

Author: Luke Zhang

.github/workflows/lint_0.yml

@@ -89,6 +89,25 @@ jobs:
       - name: Run tests
         run: tox -e lint-instrumentation-google-genai
 
+  lint-instrumentation-mcp:
+    name: instrumentation-mcp
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - name: Checkout repo @ SHA - ${{ github.sha }}
+        uses: actions/checkout@v4
+
+      - name: Set up Python 3.13
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.13"
+
+      - name: Install tox
+        run: pip install tox-uv
+
+      - name: Run tests
+        run: tox -e lint-instrumentation-mcp
+
   lint-resource-detector-containerid:
     name: resource-detector-containerid
     runs-on: ubuntu-latest