HN Explorer

Ask HN: Operational Peter Principle?

5 points by newobj ↗ HN
Phenomenon I've observed over the years, w/r/t to backend development/devops:

If you fix all the bugs that could possibly page you, then, by definition, you will only get paged by things you can't fix ("the network", undersea cable cut, data center explosions, etc).

This phenomenon feels like it needs a name, but I'm not aware of one. Is there one? What would you name it?

4 comments

[ 4.2 ms ] story [ 16.8 ms ] thread
wmf
This is not entirely true, because there's an infinitely long tail of obscure bugs. Somebody from Google observed that they mostly have really crazy bugs like "cores that don't count".
nerdponx
I think it becomes increasingly true for projects that are stable and long-lived. Maybe it's even an indicator for such projects!

If your project appears to be converging towards "no problems except those entirely outside my control" then it's a sign that your project is stable and possibly nearing feature-completion. Maybe convergence is truly asymptotic and you never fully get there, but the bugs and feature changes get smaller and smaller over time.

Not all projects can or should be "stable" in this sense, but it is an interesting indicator.

GianFabien
Contrary to the Peter Principle as applied to managements, what you describe is progressively higher levels of competence. It's like asymptotically tending to target SLA.
warrenm
>If you fix all the bugs that could possibly page you

Then you need to either:

- add more features

- find bugs that "belong" to others, and work on them

- go work from beach

Or maybe all three :)