Worked on observability and reliability improvements across dragonflydb/dragonfly and apache/druid repositories. In dragonflydb/dragonfly, refactored Prometheus metrics to separate command count from duration, enabling more granular monitoring and clearer SLA tracking, while enhancing code documentation for maintainability. In apache/druid, addressed a Kubernetes deployment issue by updating the peon container startup script to correctly source JVM options from jvm.config, reducing misconfiguration risk and improving pod stability. Leveraged Java, Shell scripting, and Kubernetes expertise to implement robust testing and ensure compatibility with existing extensions. The work focused on system programming, metrics, and operational reliability in distributed environments.