Worked on the m1a2st/kafka repository to enhance the reliability and observability of tiered storage operations, specifically addressing copy lag metric visibility during failure scenarios. Focused on backend development using Java, the work involved fixing a critical bug to ensure that RemoteCopyLagBytes and RemoteCopyLagSegments metrics are consistently emitted even when exceptions occur during segment copy processes. This technical approach improved monitoring by reducing blind spots in lag detection, supporting faster incident response and better capacity planning. Emphasized robust unit testing to validate the solution, aligning the system’s monitoring capabilities with service level objectives for copy lag visibility and operational transparency.