Worked on observability and reliability improvements in the pingcap/tidb and pingcap/docs repositories, focusing on backend development with Go, Kubernetes, and Prometheus. Enhanced the TiDB Server Status panel by deduplicating metrics per instance and aggregating with max functions to reduce monitoring noise in Kubernetes environments. Introduced new statement summary window metrics to improve visibility into execution patterns and eviction events. Improved slow query traceability by logging session connection attributes and updating related documentation in Markdown and JSON. Restored test statistics lease configuration to ensure test reliability, supporting more accurate performance analysis and capacity planning for database management workflows.