FailViz: A Tool for Visualizing Fault Injection Experiments in Distributed Systems

Abstract

The analysis of fault injection experiments can be a cumbersome task. These experiments can generate large volumes of data (e.g., message traces), which a human analyst needs to inspect to understand the behavior of the system under failure. This paper introduces the FailViz tool for visualizing fault injection experiments, which points out relevant events for interpreting the failures. We also present a motivating example in the context of OpenStack, and point out future research directions.