Worked on the elastic/beats repository to address a bug affecting AutoOps latency metric accuracy. Focused on backend development using Go, the developer updated the metric calculation logic to clamp autoops_es_latency_in_millis values, ensuring single-sample latency could not exceed the wall-clock time between samples. This adjustment eliminated misreported latency spikes, improving the reliability of performance dashboards and supporting more accurate SLA tracking. The work included adding comprehensive tests to cover edge cases, which strengthened observability and reduced the risk of future regressions. Emphasis on testing and metric reliability contributed to faster incident diagnosis and more dependable operational monitoring.