Worked on enhancing observability for the Watermark Pod Autoscaler in the DataDog/watermarkpodautoscaler repository by developing a feature that instruments autoscaler decisions with Prometheus gauge metrics. Focused on backend development using Go and Kubernetes, the work involved updating the controller logic to emit metrics reflecting autoscaler conditions such as able_to_scale and scaling_limited. Comprehensive tests were added and updated to validate metric emission and ensure reliability of the observability improvements. This approach enables proactive detection of scaling issues and streamlines troubleshooting in production environments, providing teams with actionable insights into autoscaler behavior and improving operational visibility for Kubernetes workloads.