OpenStack Summit May 2015 Vancouver

A key feature when monitoring and debugging any Cloud infrastructure is to provide the ability to trace, track, and collate all the individual, discrete steps that compose an event. A typical resource action in OpenStack is often a combination of smaller tasks -- which given the distributed nature of OpenStack -- can fail at unpredictable points in the workflow. By collecting the appropriate events, operators can view all events within Ceilometer, filter on a failed action and trace back the history of related events to spot anomalies or errors.

In this talk, we provide an overview of the recent enhancements made in Ceilometer to support the collection of event notifications from OpenStack services. We will describe: how events are processed, transformed and stored in Ceilometer; how you can derive metrics from events; and how it’s possible to track the events of a resource and analyse where errors occur.