When we work on large-scale projects, we tend to face more challenges and more complexity. That is a familiar process in development companies. A strategic look at each stage can change the quality of every delivery.
A successful project depends on several factors: sound planning, a clear scope, a method for the development work, follow-through on the project, and automation of the processes that help it meet every need. Automation is the subject of this article.
01Did you say automation? That is a job for DevOps.
DevOps can be summarized this way: “develop solutions that increase productivity, accuracy, and security in the company’s everyday tasks.” Those tasks are not limited to development and infrastructure. DevOps also carries a strong cultural responsibility inside companies, working on broader solutions with a view of flows, integration between departments, and the social interactions that happen every day, or that could happen.
Take a simple example. Every company has an onboarding flow for new employees, or should have one. In that flow, if one party fails or is late creating access for the new colleague because an emergency came up, it hurts the first days of that new team member, who can spend hours waiting for access on day one. A poor experience.
In that case, DevOps works to build a solution that gives autonomy to the person running onboarding, so they can create the required access simply and quickly.
The DevOps team can provide that autonomy in several ways, including integration with bots in the company chat. That can remove the problem and leave a strong first impression for every new colleague.
In the example above, DevOps can be essential so a flow already defined by other departments actually produces a positive result.
02Why does automation matter in large, highly complex projects?
When we talk about automation, we mean turning repetitive tasks into solutions that technology runs automatically, whenever a specific action happens or whenever someone requests it.
Why does automating matter? There are a few reasons. Large projects involve continuous delivery. The team is always producing and handing work over so the test team can validate it.
Once a task is ready, for example, it needs to be published to a test environment. Usually that is a manual task that depends on someone in leadership or on the infrastructure team. It needs a manual action before it can continue.
Manual actions are more exposed to failure than automated ones. Continuing the example, imagine that publishing the task means running a set of commands, and the person forgets a specific one. The test team then finds a series of errors that, in the end, are not a problem in the task itself. They come from a failure in the publication. That is a lot of pain.
In short, all of those problems would keep increasing the cycle time of the task, the time it takes to finish the task from the moment it starts.
Automating, then, means speeding processes, raising productivity, lifting the success rate of execution, and shortening the cycle time of tasks, which in turn shortens the cycle time of the whole project.
Publishing tasks is only one example of what can be automated. Automatic code analysis, test automation, and commands created by bots are other examples from a wide set of possibilities available to teams that aim for quality and high productivity.
The success of a project can depend on joint action between DevOps and the rest of the company, whether that is the development team, the test team, the Project Owners, or the Project Managers. It is DevOps’s responsibility to stay aligned with everyone, so it can understand the biggest difficulties and bring the right solutions for the moment.
Hugo Aguiar, responsible for DevOps at DIGITAL HUB. He holds a degree in Systems Analysis and Development from UTFPR-CP, has worked in the web development market for more than nine years, and has more than two years of experience as a DevOps professional.
Article originally published on E-Commerce Brasil.
