Worked on the stfc/SCD-OpenStack-Utils repository to enhance the reliability of Grafana dashboards during long-running Prometheus queries. Addressed persistent 504 Gateway Timeout errors by tuning HTTP timeouts in both the PrometheusRemoteWrite datasource and HAProxy, ensuring dashboards such as VM Power and Carbon remained stable when querying extended data ranges. The solution involved configuration management using YAML and Jinja, applying DevOps practices to optimize monitoring infrastructure. By carefully adjusting timeout parameters, the work resolved dashboard stability issues without introducing new features, demonstrating a focused approach to operational reliability and system tuning within a complex monitoring environment over the course of one month.