Reading: “Does the GitOps Emperor have no clothes?”

Some interesting points about the GitOps topic, where the author criticises some of the most popular benefits of this way of managing your environments.

Does the GitOps Emperor Have No Clothes?

I cannot deny I like some parts of this paradigm, where the state of your system is expressed in Git, it has a history that makes progress through known things like PRs and code reviews, and the current state of your environment is reconciled with this desired state.

However, I admit the difficulty of detecting drifts in the desired state, as well as bringing back to Git some things that happen dynamically in your environment. How much inmutable can your environment be? I like the quote from the Google SRE book: “70% of outages are due to changes in a live system”.