In August 2025, James Perng enhanced observability within the pytorch-labs/monarch repository by integrating tracing more deeply into the Monarch actor system. He refactored the telemetry layer, replacing manual enter and exit span calls with a context manager approach in Python, and ensured that actor IDs were consistently tied to tracing spans. This work leveraged his expertise in distributed systems, OpenTelemetry, and Python to deliver more granular and accurate performance monitoring. The changes improved diagnostics and maintainability, laying a foundation for future instrumentation while addressing the need for precise attribution and correlation of telemetry data across the actor system.