Worked on the open-telemetry/opentelemetry-collector-contrib repository to enhance observability for the cumulativetodelta processor by implementing three internal telemetry metrics. Leveraging Go and backend development skills, introduced metrics to track converted datapoints, dropped datapoints with reasons, and the number of streams currently monitored. The solution used an O(1) callback design with atomic counters to ensure minimal performance overhead as state scales. These metrics provide comprehensive visibility into data processing, facilitate debugging, and support reliability and capacity planning. The work focused on monitoring and telemetry, resulting in a single feature commit that improved end-to-end insight into processor behavior.