Designing a Unified Cloud Log Analytics Platform
Maoyuan Sun, Gregorio Convertino, Mark C. Detweiler · 2016
Cloud logs offer unprecedented opportunities to understand user behaviors in cloud applications and advance practices in engineering, user experience, and business. Currently, different stakeholders with relevant role-based questions use different tools for log analysis, but they share the goal of improving the same application and need to share log analyses to make decisions about application improvements. We propose a unified cloud log analytics platform. Designing such a platform is challenging because it needs to not only draw on concepts shared among different stakeholders, but also to support role-specific analyses. In this paper, we propose a conceptual framework for log analysis and show its usefulness. Then we present the design and evaluation of a prototype that implements the framework and applies visualizations to allow different non-technical stakeholders to perform log analytics. We conclude with lessons learned to inform future work.