Will DevOps steal my job?

Are you worried automation will replace people in the workplace? You may be right, but here's why that's not a bad thing.
343 readers like this.
A bunch of question marks

Opensource.com

It's a common fear: Will DevOps be the end of my job? After all, DevOps means developers doing operations, right? DevOps is automation. What if I automate myself out of a job? Do continuous delivery and containers mean operations staff are obsolete? DevOps is all about coding: infrastructure-as-code and testing-as-code and this-or-that-as-code. What if I don’t have the skill set to be a part of this?

DevOps is a looming change, disruptive in the field, with seemingly fanatical followers talking about changing the world with the Three Ways—the three underpinnings of DevOps—and the tearing down of walls. It can all be overwhelming. So what’s it going to be—is DevOps going to steal my job?

The first fear: I'm not needed

As developers managing the entire lifecycle of an application, it's all too easy to get caught up in the idea of DevOps. Containers are probably a big contributing factor to this line of thought. When containers exploded onto the scene, they were touted as a way for developers to build, test, and deploy their code all-in-one. What role does DevOps leave for the operations team, or testing, or QA?

This stems from a misunderstanding of the principles of DevOps. The first principle of DevOps, or the First Way, is Systems Thinking, or placing emphasis on a holistic approach to managing and understanding the whole lifecycle of an application or service. This does not mean that the developers of the application learn and manage the whole process. Rather, it is the collaboration of talented and skilled individuals to ensure success as a whole. To make developers solely responsible for the process is practically the extreme opposite of this tenant—essentially the enshrining of a single silo with the importance of the entire lifecycle.

There is a place for specialization in DevOps. Just as the classically educated software engineer with knowledge of linear regression and binary search is wasted writing Ansible playbooks and Docker files, the highly skilled sysadmin with the knowledge of how to secure a system and optimize database performance is wasted writing CSS and designing user flows. The most effective group to write, test, and maintain an application is a cross-discipline, functional team of people with diverse skill sets and backgrounds.

The second fear: My job will be automated

Accurate or not, DevOps can sometimes be seen as a synonym for automation. What work is left for operations staff and testing teams when automated builds, testing, deployment, monitoring, and notifications are a huge part of the application lifecycle? This focus on automation can be partially related to the Second Way: Amplify Feedback Loops. This second tenant of DevOps deals with prioritizing quick feedback between teams in the opposite direction an application takes to deployment—from monitoring and maintaining to deployment, testing, development, etc., and the emphasis to make the feedback important and actionable. While the Second Way is not specifically related to automation, many of the automation tools teams use within their deployment pipelines facilitate quick notification and quick action, or course-correction based on feedback in support of this tenant. Traditionally done by humans, it is easy to understand why a focus on automation might lead to anxiety about the future of one's job.

Automation is just a tool, not a replacement for people. Smart people trapped doing the same things over and over, pushing the big red George Jetson button are a wasted, untapped wealth of intelligence and creativity. Automation of the drudgery of daily work means more time to spend solving real problems and coming up with creative solutions. Humans are needed to figure out the "how and why;" computers can handle the "copy and paste."

There will be no end of repetitive, predictable things to automate, and automation frees teams to focus on higher-order tasks in their field. Monitoring teams, no longer spending all their time configuring alerts or managing trending configuration, can start to focus on predicting alarms, correlating statistics, and creating proactive solutions. Systems administrators, freed of scheduled patching or server configuration, can spend time focusing on fleet management, performance, and scaling. Unlike the striking images of factory floors and assembly lines totally devoid of humans, automated tasks in the DevOps world mean humans can focus on creative, rewarding tasks instead of mind-numbing drudgery.

The third fear: I do not have the skillset for this

"How am I going to keep up with this? I don’t know how to automate. Everything is code now—do I have to be a developer and write code for a living to work in DevOps?" The third fear is ultimately a fear of self-confidence. As the culture changes, yes, teams will be asked to change along with it, and some may fear they lack the skills to perform what their jobs will become.

Most folks, however, are probably already closer than they think. What is the Dockerfile, or configuration management like Puppet or Ansible, but environment as code? System administrators already write shell scripts and Python programs to handle repetitive tasks for them. It's hardly a stretch to learn a little more and begin using some of the tools already at their disposal to solve more problems—orchestration, deployment, maintenance-as-code—especially when freed from the drudgery of manual tasks to focus on growth.

The answer to this fear lies in the third tenant of DevOps, the Third Way: A Culture of Continual Experimentation and Learning. The ability to try and fail and learn from mistakes without blame is a major factor in creating ever-more creative solutions. The Third Way is empowered by the first two ways—allowing for for quick detection of and repair of problems, and just as the developer is free to try and learn, other teams are as well. Operations teams that have never used configuration management or written programs to automate infrastructure provisioning are free to try and learn. Testing and QA teams are free to implement new testing pipelines and automate approval and release processes. In a culture that embraces learning and growing, everyone has the freedom to acquire the skills they need to succeed at and enjoy their job.

Conclusion

Any disruptive practice or change in an industry can create fear or uncertainty, and DevOps is no exception. A concern for one's job is a reasonable response to the hundreds of articles and presentations enumerating the countless practices and technologies seemingly dedicated to empowering developers to take responsibility for every aspect of the industry.

In truth, however, DevOps is "a cross-disciplinary community of practice dedicated to the study of building, evolving, and operating rapidly changing resilient systems at scale." DevOps means the end of silos, but not specialization. It is the delegation of drudgery to automated systems, freeing you to do what people do best: think and imagine. And if you're motivated to learn and grow, there will be no end of opportunities to solve new and challenging problems.

Will DevOps take away your job? Yes, but it will give you a better one.

Chris Collins
Chris Collins is an SRE at Red Hat and an OpenSource.com Correspondent with a passion for automation, container orchestration and the ecosystems around them, and likes to recreate enterprise-grade technology at home for fun.

2 Comments

The challenge for automation, I think, is to design systems that work WITH people, in the way that a very smart colleague might. Even machines have a perspective in the way that they come to decisions, and maybe their decisions ought to be framed as advice rather than The Answer.
With that in mind, a system should not be designed not just to supply answers, but also answer questions. For example, "Why is that the best answer? How did you come up with that conclusion? Can you give me some background on that?" If nothing else, this might help the humans involved ask better questions.

Every repetitive task will be replaced by automation. One needs to get a clue and be ready!

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution-Share Alike 4.0 International License.