Most businesses running their own software reach the same crossroads. Something breaks, it takes too long to fix, and someone says: maybe we need a DevOps person.
Usually the answer is no — but not because the work is unnecessary.
The work is real
Servers need patching. Certificates need renewing and verifying. Backups need testing, not just running. Somebody needs to notice when disk fills, when a queue stalls, when response times drift. None of it is glamorous and none of it is optional.
Why the hire usually does not work
The trouble is that the work is steady but small. For most businesses it is a few hours a month, spread unpredictably, occasionally urgent at inconvenient hours.
That shape does not match a full-time role. A capable engineer hired for it will spend most of their time on other things, and the infrastructure work will get squeezed out by whatever is louder that week. The gaps stay invisible until something breaks — which is exactly when the hire was supposed to help.
Part-time and freelance arrangements have the same failure mode from a different direction: the work is only done when somebody remembers to ask.
When a hire genuinely is the answer
- You have enough services that the work is a full week, every week.
- Your infrastructure is unusual enough that no standard offering fits.
- Deployment complexity, not maintenance, is your bottleneck — that is a different job with the same title.
- Compliance requires that named staff hold the access.
When it is not
If the honest total is a few hours a month, you do not need a person. You need the work to happen reliably without anyone remembering to schedule it.
That is what a subscription buys: not somebody's time, but the guarantee that the boring work happens on a schedule whether or not anyone is thinking about it.
The test is simple. Write down everything that needs doing monthly to keep your systems healthy. If that list fits in a day, hiring for it is the expensive answer.