Using Git to store the state of your system means you have access to an audit log. This log contains a description of all the changes introduced. If your task tracking system integrates with Git (which is a good practice), you can typically tell which business feature is related to the system's change.
With GitOps, there is less need to allow manual access to the nodes or the entire cluster, which reduces the chance of accidental errors originating from running an invalid command. Those random errors that get into the system are easily fixed by using Git's powerful revert feature.
Recovery from a severe disaster (such as losing the entire control plane) is also a lot easier. All it requires is setting up a new clean cluster, installing a GitOps operator there, and pointing it to the repository with your configuration. After a short while, you have an exact replica of your previous production system, all without manual intervention.