Authors: Hema Latha Boddupally
Abstract: This study examines how telemetry driven code analytics can be systematically leveraged to identify recurrent defect structures in complex software systems, addressing a persistent challenge in software quality engineering where failures reappear across releases despite localized fixes. The research focuses on the problem of distinguishing isolated faults from structurally recurring defects by using runtime telemetry as a primary analytical signal rather than relying solely on static inspection or post-incident reports. The study adopts a mixed methodological approach combining empirical analysis of telemetry artifacts, structured feature engineering, and quantitative pattern detection techniques to uncover repeatable defect signatures that span code, configuration, and execution behavior. Findings demonstrate that telemetry-informed analysis enables earlier recognition of defect recurrence, improves diagnostic consistency, and strengthens the linkage between observed failures and underlying code structures. The proposed approach introduces a coherent analytical framework that integrates telemetry normalization, defect signature extraction, and code level attribution, offering a novel contribution to defect analysis practices. From a strategic perspective, the study contributes to software engineering research by reframing defect detection as a signal driven, evidence based process grounded in runtime behavior. Practically, it provides guidance for engineering teams seeking to enhance reliability, reduce diagnostic effort, and institutionalize learning from recurring failures. The work holds significance for both academic inquiry and industrial practice by advancing a scalable, analytically rigorous pathway for improving long term software quality.
DOI: http://doi.org/10.5281/zenodo.18125283
