DevOps tends to be viewed through a technical lens, but the people aspect is what will dictate your success or failure.

So, how are the people challenges of DevOps more important than the technical challenges?

Recently I was reading Puppet Labs’ State of DevOps Report. Similar to VersionOne’s annual State of Agile™ survey, this report has an interesting set of conclusions about the current state of DevOps. One of the things that I’ve been considering for some time is the strange proximity of DevOps people to technical problems.

There are a lot of technical aspects to doing DevOps. One example is continuous integration. When releasing, everything needs to be integrated and checked, no matter how small. DevOps takes it one step further with continuous deployment since you need to set up acceptance tests and unit tests, and all tests have to be automated.

One of the aspects of DevOps that reading this survey reinforced for me is that the folks who do DevOps the most successfully focus on the people aspect as much as, or sometimes even more than, the technical aspect. This brings the Gerry Weinberg conundrum to mind. Gerry Weinberg, considered by some to be the grandfather of agile, said that there’s always a people problem. The technical focus of DevOps may lead you to think that there isn’t a people aspect to focus on. In reality, while the technical piece is a focus, the people problem is huge and even more important.

DevOps is about bringing the functions of operations and development together, which means you have to break some stereotypes. You have to think outside the box. Those on the DevOps team must have the will to think about each one of the development stories through the post-deployment lens. Do you have monitoring as part of what you’re doing? Do you have performance constantly tested? Do you have all aspects of your tests automated? If not, then you’re probably not going to be as successful with DevOps.

This sounds technical, but it’s actually more of a people problem. It’s employing DevOps engineers who remember to do that. It’s not asking the question of should you automate, but how will you automate this specific story? That’s huge and it’s vital. It needs to be addressed and it needs to be addressed successfully. If you don’t, then you’re not tall enough to ride the ride. It’s not worth your time and energy to claim to be DevOps if you can’t do those things. DevOps takes time, energy and nurturing. It takes the willingness for individuals to step up. And it takes the willingness for management to step back and give your people the opportunity.

There were a couple of other fascinating people issues that that survey brought to light. The number one key to success in an IT organization, according to this survey, was employee happiness. This is an aspect people need to listen to this and build on. Employee happiness is number one. It proves that even geeky programmers like me don’t derive our employee happiness only from the really cool tools we get to use and work on. It’s nice, but it’s not enough.

The other fascinating people issue was getting the whole team together to work on the issues, thus becoming a cross-functional team. Everybody is responsible for everything, and that makes a huge difference in the DevOps world, even more so than in the traditionally agile world. Teamwork is where you’re really going to see the difference between just checking things in all the time and continuous deployment.

DevOps lends itself to being viewed through a technical lens, but the people aspect is what differentiates you from success or failure. I hope I have inspired you to take a closer look at how you are balancing the people aspect of DevOps with the technical.

So, what are some other people aspects of DevOps that should be focused on?

VersionOne and State of Agile are registered trademarks of VersionOne Inc.


Join the Discussion

    • Shikha Singal

      Agile software development methodologies have become the key element of flexible and responsive software development. DevOps is very helpful in deployment since you need to set up acceptance tests and unit tests, and all tests have to be automated.

    82 − = 78