Over the past few year we've increasingly been relying on data analytics and automation to keep on track of our increasing number of CI job results. With things like elastic recheck, subunit2sql, openstack-health, and grafana we've got an increasing number of tools to track results and provide insights into what's happening in the gate. But, even with all this data capture a great deal of the analysis has to happen by hand in these systems. Either by identifying things to track or combing through the data to find points of interest. It would be great if we could automate as much of that as possible, which is an a good application for machine learning.
This session is about bootstraping a community effort to build tooling to apply machine learning techniques to our CI results analysis to provide useful insights in an automated manner.