Worked on the vast-ai/vast-cli repository to enhance observability by implementing execution time metrics for CLI requests. This involved developing instrumentation to track per-request processing times, success statuses, and batch reporting with detailed queue entry and completion timestamps. The work was carried out using Python, focusing on backend and API development, and included updating the test suite to ensure reliability and regression coverage. After deploying the metrics, a rollback was performed to maintain system stability, reflecting a careful, stability-first engineering approach. These efforts improved visibility into CLI performance, supporting data-driven troubleshooting and laying groundwork for future optimization initiatives.