Zenith

Overview

Zenith streamlines crash management by organizing logs and simplifying error categorization. It automates error classification using machine learning techniques and clusters similar crashes to help developers pinpoint issues efficiently. A dashboard provides real-time monitoring and analysis of crash data. I wanted to explore how unsupervised clustering and anomaly detection can be used to improve developer workflows, and Zenith was the result. It performed reasonably well with strong clustering, but did not generalize across OS platforms, which was an interesting observation about how modern crashes are strongly browser-based and not OS-based.