Beyond automated expert analysis - Vitrage takes the first steps into the realm of machine learning.
Vitrage is the official OpenStack project dedicated to root cause analysis (RCA) and deduced insights into the behavior of the Cloud. With the help of a graphical model representing the different relationships between system entities, and a friendly UI, Vitrage hopes to achieve its mission of organizing, analyzing and visualizing a holistic view of the system.
In this talk, we will present and discuss the first steps of augmenting Vitrage with machine learning capabilities. Until recently, Vitrage used mostly humanly-written templates to drive automatic RCA, deduced alarms & states. The next step is to start detecting those rules in the data itself, discovering connections over time from the observed history. Here we will discuss the challenge, where Vitrage is today, and what to expect in the near future in this exciting journey.
Attendees should expect to learn about some of the new and exciting steps that Vitrage is taking into the realm of Machine Learning, with the focus on discovering how faults impact system behavior and propagate throughout the cloud, generating new insights for Vitrage to use and reflect to users.
