Worked on enhancing reliability and observability in the dreadnode/sdk repository by addressing a critical issue in distributed tracing propagation. Focused on ensuring correct propagator selection when continuing traces, the solution introduced robust handling for global NoExtract scenarios by falling back to the W3C propagator. Used Python and OpenTelemetry to validate extracted context, enabling trace continuation when possible or creating new spans to maintain accurate trace graphs. This approach improved trace integrity across service boundaries and reduced debugging complexity for distributed systems. The work demonstrated depth in distributed tracing and observability, contributing to more reliable and maintainable Python-based microservices environments.