Whereas the usage of Git because the model management system shouldn’t be strictly required, it has develop into the default selection due to its ubiquity in fashionable devops workflows and its sturdy assist for collaboration and alter administration, so its place within the title has caught.
What’s the CI/CD course of?
A whole take a look at CI/CD is past the scope of this text—see the InfoWorld explainer on the topic—however we have to say a couple of phrases about CI/CD as a result of it’s on the core of how GitOps works. The steady integration half of CI/CD is enabled by model management repositories like Git: Builders could make fixed small enhancements to their codebase, reasonably than rolling out big, monolithic new variations each few months or years. The steady deployment piece is made attainable by automated programs referred to as pipelines that construct, take a look at, and deploy the brand new code to manufacturing.
Once more, we preserve speaking about code right here, and that normally summons up visions of executable code written in a programming language comparable to C or Java or JavaScript. However in GitOps, the “code” we’re managing is basically made up of configuration information. This isn’t only a minor element — it’s on the coronary heart of what GitOps does. These config information are, as we’ve stated, the “single supply of reality” describing what our system ought to seem like. They’re declarative reasonably than instructive. That implies that as a substitute of claiming “begin up ten servers,” the configuration file will merely say, “this method consists of ten servers.”
