Developed and delivered a new Application Performance Monitoring metric for the elastic/elasticsearch repository, enabling tracking of the DFS search phase duration at the shard level. This feature was implemented using Java and YAML, with a focus on backend development and performance monitoring. The approach involved instrumenting the codebase to capture granular timing data, ensuring consistency with existing APM dashboards and metric naming standards. By providing detailed shard-level insights, the work enhanced observability and supported faster root-cause analysis, while also informing capacity planning and SLA adherence. Unit testing was applied to validate the metric’s accuracy and maintain integration reliability within the system.