My Ops Team Is a Pile of Cron Jobs

For years my job was running an operations team at a healthcare company. A real one, with people. They watched the dashboards, caught the problems before anyone else noticed, provisioned the new hires, reconciled what the systems claimed against what was actually true, and told leadership every week what mattered. Boring, essential work. The kind nobody thanks you for until the week you stop doing it.

Now I run businesses by myself—a card shop, a couple of content sites, the books behind them—and I have no team at all. So I did the only thing twenty years of infrastructure prepared me to do: I rebuilt the team out of scheduled jobs.

What an Ops Team Actually Does

Strip away the org chart and an ops team does five things. It watches the systems and notices when one goes sideways. It keeps the records honest—making sure the numbers in one place match the numbers in another. It stands things up and tears them down: new accounts, new devices, new listings. It handles the incident at 2am. And it reports up, so the person in charge knows what's going on without having to go digging.

None of those are glamorous, and none of them require a human in particular. They require something reliable that runs on a schedule and doesn't get bored. A person is a very expensive way to do a thing that happens the same way every fifteen minutes.

Every Role Becomes a Job

So I mapped each one onto a job that runs on its own.

Monitoring. My card business runs dozens of scheduled jobs that sync channels, check their own health, and refresh prices. They are the team member whose entire job is to stare at the board and raise a hand when something looks off— except they never blink and they work every night.

Reconciliation. The finance system behind the shop is nineteen scheduled jobs pulling five sales channels into one database and reconciling the books to zero. A human bookkeeper does this once a month and dreads it. The jobs do it continuously, so drift gets caught the day it happens instead of the following April.

Provisioning. When a card sells, one operation records the sale and delists that card everywhere else. When I list a new one, a photo becomes a finished listing. The stand-it-up, tear-it-down work that used to be somebody's afternoon is now an event handler that fires and moves on.

Reporting. Every morning one short email tells me what matters—at most six sentences, each a number and the reason it surfaced. That's the weekly leadership update, shrunk to a daily habit and handed to an audience of one. Nobody builds a slide deck. The slide deck builds itself and keeps it to six lines.

The Night Shift Is the Hard Part

Anyone can schedule a job. The thing that makes it an ops team instead of a pile of scripts is how it behaves when something breaks—because something always breaks.

A good ops team has a rule: you don't page a human for a problem that fixes itself. So my jobs assume failure. A browser session dies, the job relaunches it. A transient error, it retries once, quietly. It only reaches for me after it's failed twice in a row, which means a network blip never becomes a 2am notification. An alert you've learned to ignore is worse than no alert, because now you've trained yourself to sleep through the one that matters.

The other half is that the jobs heal instead of sync. The reconciliation loop doesn't nudge state from one system to another and hope they stay in step—it throws away the copy and rebuilds it from the source every time it runs. A job like that can't drift, and if it crashes, the next run makes everything right again. That is the actual definition of an ops team you can trust: the one that cleans up after itself before you've even noticed the mess.

The Team That Scales to Zero

The strange part is how little of this was new. The healthcare job taught me the whole playbook—identity as the source of truth, least privilege, document the decision, automate before the manual process calcifies into debt. I just pointed the same playbook at businesses too small to hire anyone.

And that's the quiet advantage. A real ops team has a headcount, a budget, and a bad week when two people are out sick. Mine has none of those. It costs almost nothing, it doesn't take vacation, and it gets a little bigger every time I write one more job. A one-person business isn't understaffed. It's just staffed by things that run on a schedule.

I still think of it as a team. It's just that every member is a cron entry, and not one of them has ever asked me for a raise.