Git isn’t just for developers.
It’s a time machine for anything that can be written down.
When you self-host or build systems solo, half the job is remembering what changed and why.
Without history, you’re debugging blind.
Everything as Text
Anything you can put in plain text can live in version control:
- Web server configs
- Backup scripts
- Hugo site templates and content
- Firewall rules
Text has the advantage of being diffable. You can see the change, not just the result.
Commits as Documentation
No one reads a separate changelog. But they will read commit messages:
- “Switched cron backup from 2 a.m. to 3 a.m. after clash with updates”
- “Added rate limit to nginx after brute-force attempts”
History plus reasoning — future you will thank you.
Rollback Is Cheaper Than Guesswork
When something breaks, guessing which file changed is a time sink.
Git gives you:
- A snapshot of the last known good state
- The ability to roll back instantly
- A forensic trail when you need to explain what happened
Cheap insurance with zero ongoing cost.
GitOps Without the Buzzword
Big companies call this “GitOps”.
For one-person ops, it’s just common sense:
- Repo on your machine
- Push to a private remote (even a USB drive counts)
- Pull changes before you start messing with configs
No need for pipelines and Kubernetes if you don’t have the scale.
Just version the text and move on.
Stop Trusting Memory
The human brain is bad at state management.
Version control externalises that memory.
Treat all infrastructure like code: logged, diffable, reversible.
Git works. Use it.