Training: June 20–21, 2016
Tutorials: June 21, 2016
Keynotes & Sessions: June 22–23, 2016
Santa Clara, CA

Under the covers: Capturing application errors, logs, and metrics

Bryan Liles (Heptio)
1:30pm–3:00pm Tuesday, 06/21/2016
Measuring the right things, Performance for the people
Location: Ballroom GH Level: Intermediate
Average rating: ***..
(3.38, 13 ratings)

Prerequisite knowledge

Attendees should be familiar with web topics around application development or operations and have experience with tools like cURL and SSH. Application programming abilities are not necessary.

Materials or downloads needed in advance

Attendees need a laptop with Internet access and the ability to run tools like SSH and cURL. DigitalOcean will provide all server resources.


This tutorial is for you because:
You work in Ops and need to keep track of how your systems are running.

Writing an application is only half the battle. Once an application goes to production, insight is required to determine how it’s performing, and it is up to developers and operations teams to capture the data for future analysis. This data includes logs, metrics, and errors. The insights this data reveals allow developers to build more resilient systems in the future.

Bryan Liles walks attendees through scenarios that will allow them to view application metrics and logs and capture errors for further analysis. You’ll learn how to configure systems to capture relevant data, how to expose systems to logging and metrics as a practice rather than an afterthought, and how to use measurement as an integral part of your system design and operation.

We will use a Go application in this session, but the lessons apply to multiple languages.

Over the course of this workshop, attendees will:

  • Configure logging for a web service that performs multiple operations to see real-time feedback
  • Create multiple types of metrics using Prometheus and learn the positives and negatives of push and pull systems
  • Configure a system to capture error logs that can be used to inform the production teams of regressions that are occurring inside of their applications
  • Examine a system and reflect on the insights revealed and how they help improve future iterations
Photo of Bryan Liles

Bryan Liles


Bryan Liles is an engineer at Heptio. When he is not writing software to help move teams to Kubernetes, he gets to speak at conferences on topics ranging from machine learning to building the next generation of developers. In his free time, Bryan races cars in straight lines and around turns and builds robots and devices.

Comments on this page are now closed.


Picture of Bryan Liles
Bryan Liles
04/29/2016 7:15am PDT


There will be no Go development in this project. I will most likely use a Go app as a part of the demonstration, but it won’t require any knowledge above knowing what a source file is, and how to edit it. In this session, there won’t be any advanced requirements, but if you wanted to familiarize yourself with and, that could be a bonus.

Picture of Matt Brender
Matt Brender
04/29/2016 4:47am PDT

Are we going to dig into Go development in this one or use a pre-created binary? Are there any packages worth familiarizing ourselves with before attending? Thanks!