Doing DevOps in an Interconnected World
Doing DevOps would be simple if you had total control over the tools, code and infrastructure that your company used. But the fact is, you don’t. Instead, we live in a world where most CI/CD chains connect in some way to third-party resources or processes.
In other words, doing DevOps effectively means integrating with entities that are external to your organization. Here’s a look at what this entails in practice and why it’s so important.
The Interconnected World of DevOps
Conversations about DevOps often assume that all DevOps processes happen internally. In reality, many software delivery pipelines depend on third-party resources and services. Consider the following examples of ways companies could make their software delivery chains dependent in part upon third-party agendas:
They integrate open source code into their codebases and therefore depend on an open source community to write some of their code.
They maintain relationships with partners and need to be sure their release schedules and toolchains are compatible with their partners’ needs.
They need to make sure the applications they deliver work on different operating systems, browsers and hardware devices that are controlled by third parties.
They are subject to regulatory rules handed down by third parties and must ensure their delivery chains meet those regulatory requirements.
Their applications depend on APIs that are designed and updated by third parties.
In these ways and more, the CI/CD pipelines organizations build to do DevOps need to be able to integrate with, and adapt to, decisions made by third parties. If they don’t, the organizations risk problems such as broken code, the loss of partners and regulatory compliance issues.
Succeeding at Interconnected DevOps
This is why it’s critical to design CI/CD pipelines and other DevOps processes in ways that make them mesh with third-party resources and services. In practice, this means:
Adopting tools that are flexible and have broad coverage. You don’t want to tell a partner you can no longer work with their code because your CI server doesn’t support their new programming language.
Use branching in the CI/CD process. Multiple branches give you more flexibility to support different target environments. They also help you react to sudden changes made by third parties without having to overhaul your entire CI/CD process.
Make sure your collaboration and communication channels integrate with those of third parties on whom you depend. If you use open source code in your application, for example, add the open source project’s security announcements to your Slack channel so you’ll be notified right away when a vulnerability arises.
Choose your integrations wisely. Just because you can integrate a third-party API into your application, for example, doesn’t mean you should. Too many dependencies put the efficiency of your CI/CD pipeline at risk.
No matter the details of your CI/CD chain, the most important thing is simply to recognize that we live in a very interconnected world and most organizations don’t have total control over their software delivery processes. That’s just a fact of modern DevOps. But with the right planning and strategies, you can ensure that third-party dependencies add value to your software delivery pipeline rather than creating unnecessary risk and overhead.