The "sound" of performance monitoring data (sponsored by Riverbed)

Description

Over the past 15 years, Jon Hodgson has helped companies all over the world solve performance issues in their most critical applications. One recurring theme Jon has encountered is that inadequate monitoring data either masks or misrepresents the true symptom or root cause, leading their troubleshooters on a futile quest that never leads to a resolution. One of the main culprits behind this is aliasing, artifacts and errors which occur when a signal is improperly sampled. You’ve likely encountered this phenomenon in low-quality digital audio, image, and video files, which are distorted when compared to their original source.

The digital media industry leverages fundamental digital signal processing principles to ensure that analog events are captured with precise fidelity, yet in the field of performance monitoring, these principles seem to be largely absent in many tools, which then imprecisely capture the performance data we rely on.

Using sound as a surrogate for monitoring data, Jon offers an overview of digital signal processing concepts and explores analogous real-world cases where performance data effectively “lied” to the troubleshooter, suggesting that overloaded resources weren’t and vice versa. You’ll learn how to avoid these pitfalls and determine if the data from your monitoring tools is sufficient to detect and quantify certain common problems and how to read between the lines to identify when your data can (and cannot) be trusted. Along the way, Jon explains how other fields have overcome these challenges and how some of their unconventional techniques can be applied to DevOps methodologies.

This session is sponsored by Riverbed.