fix: clear leaked running loop in MCP client background thread

[mkmeral]

Description

When opentelemetry-instrumentation-threading is active — the documented pattern for AgentCore observability — the ThreadingInstrumentor wraps Thread.run() to propagate trace context. Combined with contextvars.copy_context() in MCPClient.start(), this leaks the parent thread's running event loop reference into the background thread. When _background_task then calls run_until_complete(), asyncio sees a "running" loop already set and raises RuntimeError: Cannot run the event loop while another loop is running.

This adds asyncio._set_running_loop(None) at the top of _background_task() to clear leaked loop state before creating the new event loop. This is the same pattern the bedrock-agentcore team shipped in aws/bedrock-agentcore-sdk-python#405.

Confirmed on SDK versions 1.28.0 and 1.35.0.

Related Issues

N/A

Documentation PR

N/A — no documentation changes needed.

Type of Change

Bug fix

Testing

How have you tested the change? Verify that the changes do not break functionality or introduce warnings in consuming repositories: agents-docs, agents-tools, agents-cli

• I ran hatch run prepare

Added a unit test that simulates the OTEL leak by setting a running loop before _background_task executes, verifying the method clears it and completes successfully.

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[github-actions]

Assessment: Comment

Well-scoped bug fix that addresses a real issue with OpenTelemetry's ThreadingInstrumentor leaking running event loop state into MCP client background threads. The fix is minimal, the approach is validated by the bedrock-agentcore team, and the test effectively simulates the failure scenario.

Review Details

• Private API usage: asyncio._set_running_loop(None) is a CPython internal — stable and correct for this use case, but worth a brief note for future maintainability.
• Test import style: import asyncio should be at the module top-level per repository coding standards.

Clean fix with good context in both the PR description and code comments.