Missing telemetry API/provider causing misleading exception

[onmete]

🤔 What is the technical debt you think should be addressed?

Currently, if the telemetry API/provider is not defined in the run.yaml configuration, the llama-stack fails with an UnboundLocalError in agent_instance.py (turn_id variable not associated with a value). This failure path is misleading, as the root cause is simply that telemetry is required but not configured. The code does not provide a clear error message indicating that telemetry is mandatory, which makes debugging more difficult.

Error:

  File "<project-path>.venv/lib/python3.12/site-packages/llama_stack/providers/inline/agents/meta_reference/agent_instance.py", line 191, in create_and_execute_turn
    async for chunk in self._run_turn(request, turn_id):
                                               ^^^^^^^
UnboundLocalError: cannot access local variable 'turn_id' where it is not associated with a value

💡 What is the benefit of addressing this technical debt?

Improving error handling to explicitly indicate that telemetry is required would:

• Prevent developer confusion when debugging missing configurations.
• Reduce wasted time interpreting misleading stack traces.
• Make the system's requirements clearer and easier to document.
• Lay the groundwork for potentially making telemetry optional in the future, or at least providing graceful fallback behavior.

Other thoughts

No response