I am a firm believer in learning on the job. I believe that the lessons learned best are those learned through experience and action. While I have never been in a company as it transitioned to DevOps, I decided to take a note from J. Wolfgang Goerlich who spent a year making the transition and “making it up as he went along”. Without further ado…
1. What is DevOps?
First, let’s define our terms. Goerlich describes DevOps as a system for taking the best of development and the best of operations and making something that’s better than the sum of its parts. At the core of it, it’s about making infrastructure more like code in such a way that you can actually “version” your infrastructure. But virtualization isn’t the key thing, he says: automation is. Virtualization is just a way to support automation.
Developers who want to get started with DevOps start by reaching out to ops and getting themselves on the call list so they can see what sort of issues ops has to deal with every day. If you’re in infrastructure, reach out to development and get in on some build meetings. Learn about how that process works and the pressures they are under.
3. What You Need to Measure
Metrics are an important part of measuring success, not just for making the case for DevOps to management but for making sure that you really are doing it right. Goerlich suggests three specific areas to measure:
- Implemented changes: How frequently do you change? You don’t have to be making changes every 11 seconds like Amazon.com does but you do want to keep track of how many changes you’re pushing out so that you can prove that you’re doing more.
- Successful changes vs. failed changes How many of these changes had to be rolled back? Track this so that you can prove that you’re making fewer mistakes.
- Help desk tickets related to changes vs. help desk tickets not related to changes.Are you causing more problems for end users or fewer? Track this to prove that you’re not only doing more, but you’re doing it better and helping users get their work done.
The “one team, one system” approach means viewing infrastructure through a single pane of glass.
5. Keep Your Employees Smart
Goerlich says his team spends 20% of their time on training so that they can not only meet current requirements of the company but be ready for the technological changes coming down the pike.
6. What a DevOps Superstar looks like
- Know your own area inside out.
- Don’t limit yourself to your own area. If you’re a developer, learn about infrastructure and security. If you’re in ops, learn about development.
- Take responsibility for your input.
This is only a partial version of Georlich’s lessons learned in the DevOps trenches. Click here to read the full article by Klint Finley on the DevOps Angle.