I’ve been using AI to help with commit messages for a while now. Yesterday, in a discussion with co-workers, it became clear that this may not just be a convenience feature — it’s turning into a real time saver.
That was the background for creating the new rsyslog Commit AI Assistant. It directly addresses a problem we ourselves face in daily development. True to dogfooding, we now use it internally whenever we craft a commit message — myself included.
Want to give it a try: use the rsyslog commit Assistant.
A quick side note on commit messages
Good commit messages are more than AI fodder or something nice for reviewers. They are invaluable years later, when you need to understand why a change was made. With a project like rsyslog, that can mean going back 10 or 15 years. If the message isn’t clear — and the code comments aren’t either — you’re in for a painful hunt.
Newer projects with flashy tooling may only experience that problem in a decade or so — if they still exist by then.
More than just commit messages
Alongside message generation, I added a quick code review step. It’s powered by the AI’s training data plus the patch itself. The training data only gives a rough outline of rsyslog, but in this setting that’s good enough:
- It forces the AI to think about the patch in broader context.
- It provides a fast sanity check, even without a full review.
In practice, this already saved me from launching an unnecessary CI run within the first 36 hours of using it — which not only spares time, but also energy.
Practical, not perfect — but part of AI First
The commit assistant is far from perfect, but it’s a practical everyday tool. With about an hour of prompt engineering — and some of my rsyslog know-how baked in — we got a working helper that is continuously improving as we learn from real use.
It’s a low-hanging fruit, but one that pays off immediately. And it fits nicely into our broader AI First strategy — making AI part of the everyday developer workflow, responsibly and with clear benefit.