The main difference between a classical CI/CD system and GitOps lies in the mode of operation. Most CI/CD systems are imperative: they consist of a sequence of steps to be taken in order for a pipeline to succeed.
Even the pipeline's notion is imperative as it implies an object that has an entry, a set of connections, and a sink. Some of the steps may be performed in parallel, but a process has to stop and wait for the depending step to finish whenever there is a dependency.
In GitOps, the configuration is declarative. This refers to the entire state of your system – the applications, their configuration, monitoring, and dashboards. It is all treated as code, giving it the same features as regular application code.
