Continous integration jobs can generate massive volumes of data. When a job fails, figuring out what went wrong can be a tedious process—one that involves investigating logs and discovering the root cause, often found in a fraction of the total job output.
In this workshop, we’ll show attendees how machine learning can help solve the challenge of anomaly detection in job logs and save time in finding a failure’s root cause. We’ll also demo Zuul v3 and how it can be extended to integrate log processing models. Based on a live dataset of Zuul jobs, attendees will use the log-classify project to build a machine learning model and automatically extract anomalies. The workshop organizer will provide instructions and answer any questions, but attendees are encouraged to contribute to the discussion.
Attendees should expect to learn what the challenges of anomaly detection are in continuous integration jobs and how machine learning can provide a powerful service to assist failure investigations.
The workshop will also present how the new Zuul Version 3 works and how it can be extended to integrate log proccessing models.
