DevOps without the culture is sysadmin
01 Apr 2024
Alexander III Bridge, Paris, October 2023. @babafemio
As businesses work hard to find ways to increase their efficiency and speed of delivery, they have turned to DevOps, a popular approach to software engineering and product delivery. It brings together development and operations to create a seamless process that results in faster delivery, higher quality, and better collaboration between teams. However DevOps does not always deliver on the promised and the core reason is that DevOps is not always being treated as a culture but a role or even worse a set of tasks.
A good example is the significance of the software deployment pipeline, its treatment, and associated technical debt. The pipeline plays a critical role in the software development and deployment process, and it is not uncommon for technical debt to accumulate in the pipeline. Therefore, it is essential to consider the pipeline as an integral part of the software development process and invest in maintaining it properly to maximize efficiency. To illustrate, a 30-minute delay in any phase of the pipeline would cost 2 hours of lost productivity for every 4 deployments. Therefore, addressing technical debt in the pipeline should be treated as an engineering effort and not just a ‘DevOps’ effort because fixing technical debt is a vital part of the culture of engineering excellence. In addition, the solution to addressing the debt may also lie in the code or in its architecture.
At its heart, DevOps is a culture of improvement. The goal is to continuously enhance software delivery and enjoy doing so. To improve a process, you need to measure it, and this is where monitoring comes in. In the 2-hour loss example above, the pipeline may not have started out that way. However, measuring and monitoring should trigger an alert for the entire team to swarm the problem and fix it, with ‘team’ being the keyword. This is important because a lot of times when you have someone on the team with the title of ‘DevOps Engineer,’ the task of fixing it tends to fall to them. When this happens, DevOps falls apart, and we are back to implementing good old-fashioned sysadmin.
Implementing DevOps requires a culture of collaboration, automation, monitoring, continuous improvement, and the right tools to be successful. But getting DevOps to work right is not without its challenges. Treating DevOps as a way of life rather than just a set of tasks or a role is the first and, by far, the most crucial step. Here are a few ideas on how to make DevOps work for your team.
Build a culture of collaboration
DevOps is all about bringing teams together to work towards a common goal. This requires a culture of collaboration and open communication. Encourage teams to work together, share ideas, and provide feedback. This will help to build trust and encourage innovation. The bottom line is that everyone needs to be a part of this, not just the DevOps engineer.
Monitor and measure everything
Monitoring and measuring everything is critical to the success of DevOps. This includes monitoring applications, infrastructure, and processes. Use metrics to track performance, identify issues, and measure improvements. This will help teams to continually improve and make data-driven decisions.
Focus on continuous improvement
DevOps is not a one-time project; it is a continuous process. Once the team has agreed on the key metrics to track, focus on making small, incremental improvements over time. This will help to reduce risk and ensure that teams are always working towards a common goal. If you plan to use any of these key metrics and improvements as part of performance management, ensure they are shared by the entire team.