As we approach Jenkins World, scheduled for later this month in San Francisco, we sat down with Jez Humble, author of “Continuous Delivery” and one of the best-known names in DevOps. Jez will be keynoting this year’s event, along with CloudBees CEO and founder Sacha Labourey. Here’s what he had to say.
Alan Shimel: What are you most excited for at Jenkins World?
Humble: This is my first Jenkins World! I’m most excited to meet the community and find out what people care about, what problems they’re solving, what obstacles they’re facing, and what’s making them excited or sad. I’ve spent over a decade doing and teaching continuous integration and working on tools, so I’m always looking to validate and challenge my views.
Shimel: What do you plan to cover in your keynote?
Humble: I’ll be talking about DevOps, and how it’s transforming the world of software and the businesses it powers. I’ve been working on a research project into DevOps for the last few years so there will be some hard data. I’ll also be covering the principles and practices, as well as the obstacles that teams commonly face implementing them, and how to overcome them.
Shimel: What do you hope attendees take away from your talk?
Humble: First, I want folks to understand that this transformation is real, not just hype. High performers achieve higher levels of throughput, stability, and quality, and lower costs. Second, it’s hard. We’ve been talking about continuous integration for nearly twenty years now—it’s one of the basic agile engineering practices—and still the majority of people who think they’re doing it actually aren’t. But my other message is that whatever kind of organization you’re in, whatever your technology stack, you can apply these principles today. The essence of high performance is to work on getting better every day as part of your daily work, and that’s something all of us can—and should—commit to.
Shimel: What are the major DevOps trends you’ve seen this year?
Humble: I’ve not seen anything brand new jump on the scene this year. I think that’s a good thing. We’ve had so many transformational new technologies in the last few years from containers and serverless to new tools that I think we were in danger of drowning under shiny objects and hype. I think what I’ve seen in the last year is a better understanding of the contexts in which these tools and technologies can be effectively applied.
Overall, the last few years has been pretty exciting as we’ve seen devops and continuous delivery be applied in every domain from government (I spent most of last year working at 18F in the U.S. Federal Government helping to implement a DevOps, cloud-first approach to software delivery there) to mainframes in financial services to firmware. One of the common objections to devops is the “won’t work here” attitude, and we’ve comprehensively demolished that.
Shimel: How do you expect DevOps to change in the next five years?
Humble: I’m hopeful that many of the core practices such as infrastructure-as-code, continuous delivery, disposable/immutable infrastructure and blameless post-mortems will become business-as-usual in organizations of all types and sizes. However the field is still moving pretty fast, so I think we’ll continue to see plenty of innovation both in terms of the practices, their implementation, and the tools.
Shimel: What are the three most important things DevOps professionals need to focus on to achieve DevOps success?
Humble:
- Know what success looks like in measurable terms. What are the organization-level goals we’re all working towards, and when do we want to have achieved them? Leaders should be setting clear, prioritized, measurable goals—for example, “Release every two weeks without anybody working outside normal business hours”—and then giving teams the resources and support they need to achieve them.
- Most of your problems are going to be cultural. People like to focus on tools, and tools are definitely important, but that’s not normally where the failure modes are. I recently said my #1 devops advice was to buy lunch for and actively listen to someone in the role you complain about most in your org, whether that’s the auditors, the testers, or the DBAs. Find them and work out something you can do to make their life a bit better and then do it.
- You’re never done: keep going! As I said, high performers are always working to get better. Devops and continuous delivery are implemented through working to get a little bit better every day. Most of the big lessons I learned in my career I learned not by working on exciting new technologies but by doing unglamorous grunt work using tools like CVS and bash. Find out what your team’s biggest obstacle is right now and do something to improve. Then find the next obstacle! Rinse and repeat.