Theia: Visual Signatures for Problem Diagnosis in Large Hadoop Clusters

Visualization tools play an important role in summarizing large volumes of data by revealing interesting patterns such as trends, gaps, and anomalies in the data. Users can leverage visualization tools to identify problems in their programs quickly. In this article, we present novel visualizations that help users diagnose problems in Hadoop applications. These visualizations allow users to identify problematic nodes in the cluster quickly, and distinguish between different classes of problems.