Worked on stabilizing alerting reliability in the apache/dolphinscheduler repository by addressing a bug in alert status calculation within the Alerting module. Refactored the logic for counting alert successes and failures, replacing manual tallying with a filter-based approach to improve both correctness and efficiency. Enhanced unit tests in Java to ensure the AlertSender component accurately updates alert statuses across scenarios where plugins succeed, fail, or partially succeed. Focused on backend development and alerting systems, the work improved maintainability and reliability by tightening alert status logic and expanding test coverage, contributing to a more robust alerting infrastructure for the project.