Devops culture is quickly gaining ground with companies all over the world and the demand for top notch devops talent is as high as it has ever been. With the base salary for an entry-level site reliability engineer is more than $100,000 annually, IT professionals are hurrying to make the transition into devops.
But, if you think about it, the position of devops engineer doesn’t really exist nor is it a position or a job title. Do not spend months in searching for mythical devops engineer; it does not exist in the nature. Let’s take a closer look and find out why.
From the top to the bottom
Devops does not begin with hiring. First of all, secure 100 percent buy-in from other VP-level executives to provide operations improvements and transformations within the company and, in doing so, passing along the baton to the senior executives to accept and meet new challenges. Every VP has to commit on automation, quality, and other strategic changes.
Then identify leaders within a company—people who spent at least a few years in development, software architecture, system engineering, and operations and possess the soft and leadership skills to be able to pitch and deliver things. It is OK to bring a contractor at this stage to assist you if you lack in-house expertise. Consider hands-on people versus PowerPoint slide makers because you’ll need to deliver quick wins for the business to get further buy-in.
Mentor and set higher standards for the rest of the team and for the new hires. Every engineer should know how to build software in a devops way—care about metrics and logs in production, configure Jenkins pipelines and Docker files.
Every infrastructure guy should live a developer’s life as well, supporting and automating those tools the same way as developers automate business processes. Even though devops is a rapidly growing field and “devops engineers” are in hot demand, it’s important to keep in mind that devops is not a skill of one person. Everybody in your dev team must know Linux, Docker, Docker Compose, Kubernetes, and Ansible, at least on a user level, as well understand networking and deployment architecture.
Of course, there will be people who can dedicate themselves to the infrastructure projects, but all your hires should support the idea of establishing and supporting devops culture.
Also, there is still a lot of admin or IT work in the cloud—configure IAM roles, VPN, SSO, user management, etc. It could be easily done by IT guy who learned new cloud and automation tools. But it is still important to follow a devops mindset in IT operations to prevent chaos in the cloud.
From the bottom to the top
Every product, service, API, library, and even Java function or Bash script has clients—internal or external. Engineers who deliver that software should be responsible to making users happy—provide a documentation, deployment suite and maintenance. This is true for all the internal dependencies between teams and people. So, do not hire devops engineers; hire engineers who care about the clients of their deliverables.
Extending that paradigm to the architecture and deployment levels as well as to macro level of dependencies between different teams is the way to go.
When and who to hire
Try to hire the best talent as usual. Instead of concentrating on specific tools, concentrate on candidate’s culture, background, and understanding of how things work under the hood. You will be better off finding people who understand methodologies instead of tools. It is more important that candidates comprehend the concept of CI/CD instead of asking them if they prefer Jenkins over some other automation tool. Don’t get too caught up in the exact tool chain. Rather, focus on the problem-solving skills, their ability to increase efficiency, save time, and automate manual processes. You do not want to miss out on the right candidates just because some tool is missing from their résumés.
Hire contractors to help you hire in-house and solve the chicken-and-egg problem. The sooner you’ll build the right engineering culture and devops process, the easier you’ll attract the right talent for full-time.
Use niche contractors who have templates and automated tool set to minimize time to market. Repetitive projects—such as securing an infrastructure for PCI DSS certification, building an infrastructure for Kafka-based stateful microservices and data pipelines—would benefit from having an out-of-the-box solution. But make sure the vendor does not lock you in with something proprietary and something you do not own or could not maintain.