You are cloning into a /var/www folder the purpose of which is solely to keep files that are meant to be public. Which is not always the git repo. I see that you have a reverse proxy setup, which makes only the public items on port 80. Having done so, still isn't a good setup architecturely to do things inside the /var/www folder. It is not meant to be for that, and just calls for re-inventing the wheel.

Secondly, with this approach, I wonder where is the build process? Also, there isn't a 1 second rollback step, and for rollback you are going to go through the git branching again, which is ok, just consumes time when you quickly want to rollback things. In case of Docker, you can switch the images and in case symblink deployments, it takes a second to make the rollback. If this is a personal project, then fine. But won't even pass the review on any enterprise application.

The repository https://github.com/emmettgb/EmsComputer.jl is just a static static site, which would work with just git. But git objects can become heavy upon teams of development, and is just not a scalable strategy. Even when it is a static site, and there is no build process, I still wouldn't call it a good strategy or scalable, as your article doesn't mention a strategy to save any API keys (if any).

Finally, you should add logs/.gitkeep and logs/logs.txt to .gitignore.