Old Bhopal teaches you something about cities that maps never quite capture: demolish a lane and the trade it carried does not vanish — it reappears, three streets over, under a new awning, run by the same families. Operations work behaves the same way. The pitch behind NoOps says a company can purchase and automate so much of its operational burden that hiring operators stops making sense, and plenty of businesses get close. What they actually achieve, almost every time, is a relocation. The duties pack their bags and move. This listicle is a walking tour of the seven addresses where they turn up, so that nobody mistakes a move for a disappearance.
Where NoOps Work Actually Goes: Seven Stops
The itinerary, for those who like a route card before setting out:
- the provider’s datacentre, where the heavy industry now lives;
- the internal platform team, keeper of paved roads;
- your own product engineers, carrying the pager they shipped;
- the automation code, a resident that needs feeding;
- the contracts drawer, where vendor management keeps house;
- the crisis hour, when the provider fails and duty returns;
- future you, holding every deferred decision at interest.
Stop one: the provider’s datacentre
First stop, the obvious one. Hardware, power, cooling, patching, physical security — the heavy industry of operations now belongs to whoever sells you the platform. Their engineers genuinely carry this pager, at a competence no small team could match, and the fees you pay each month are the wages of that shift. The catch hides in the contract: they carry their side of the line. The part of the operation that involves your data, your logins, and your settings never left your building.
Stop two: the platform team
In companies big enough to have one, an internal platform group becomes the new home for operational toil. They pave the roads — pipelines, templates, guardrails — so product developers can deploy without learning infrastructure from scratch. The work did not disappear when the ops department merged away; it moved upstairs and changed its job title. If you have ever heard a developer say “the platform handles that,” you have heard a relocation being described.
Stop three: your product engineers
This is the stop that surprises nobody more than the engineers themselves. The 2012-era Netflix arrangement — the one that borrowed the NoOps label — put the deploy button and the pager in the same pair of hands, and those hands belonged to developers. Veterans of that era were quick to point out that operations had been reorganised, not abolished. Today’s version reads: the person who ships the feature carries the alerts for it. Their sprint plans quietly grew a column nobody roadmapped.
Stop four: the automation code itself
Every script you write to end a manual chore becomes a small resident that needs feeding. Pipelines break when their APIs shift. Configuration drifts. Certificates expire on bank holidays with impeccable comic timing. The industry phrase is undifferentiated toil, but the human version is simpler: someone must tend the robots, and tending robots is operations wearing overalls. Teams that automate carelessly just hire a new colleague made of bash scripts.
Stop five: vendor management
Sign enough managed contracts and a new operation condenses around the paperwork: renewals, access reviews, uptime reports, the annual negotiation, the awkward question of what happens when a vendor raises prices or retires a product. Somebody has to read those contracts. Somebody has to keep the list of who holds which keys. Procurement was always operations in a suit; NoOps simply gives it a bigger wardrobe.
Stop six: whoever answers when the provider fails
Managed does not mean infallible. Regions wobble, queues back up, a status page turns orange at the worst hour, and at that moment the duty of communication — telling clients, deciding whether to fail over, choosing whether to wait — lands squarely back on your team. Reliability engineers measure this readiness in coverage hours, and the numbers are bracing: keeping a genuine day-and-night response alive takes far more people than the underlying work ever did. Teams that cannot staff that coverage have not eliminated it. They have decided to discover it live.
Stop seven: future you
The quietest relocation of all. Every duty deferred — the migration postponed, the backup never rehearsed, the exit plan never written — waits at an address called Later, and Later charges interest. A data law arriving in 2027 will make leaving a cloud provider cheaper in fees without making it cheaper in effort, and every team that skipped the exit arithmetic will meet that effort in a single, badly timed month. Future you is the most expensive operations hire you never approved.
The NoOps Tour, Honestly Concluded
Stand back from the seven stops and a pattern emerges — the same pattern a walker notices in a city that has rebuilt itself twice. Nothing was demolished. Everything moved. The operation that once sat in a server room now spreads across a provider’s shifts, a platform crew’s roadmap, your engineers’ evening alerts, a folder of scripts, a drawer of contracts, a crisis playbook, and a set of decisions you lent to your future self. The tour has an official map, if you want it: the Ops Atlas NoOps entry keeps the definitions, the history, and the maturity ladder teams use to grade how deliberately the work was moved.
That ladder is worth a sentence, because it turns the tour into a self-portrait. Its ground floor is NoOps by assertion, the memo that arrives before any evidence does. Its top floor is low-ops by design, where every stop on this tour has a name against it and nobody pretends the street has emptied. The difference between the rungs is not technology. It is whether the moving was done on purpose.
Planning Your Own Walk Through NoOps
Here is the practical close, offered in the spirit of a good local tip. Once a quarter, walk the seven stops for your own operation and write down the name living at each address. Who carries the pager for the platform? Which engineer’s evening owns the deploy? Where does the automation’s maintenance live? Who reads the contracts, and who would answer if the provider went dark midweek? Any stop with an empty name field is not a saving. It is a vacancy, and vacancies in operations are filled eventually — by whoever happens to be standing nearest when something breaks.
One warning from teams that have walked this route: the moves happen slowly, then all at once. A duty drifts to a vendor with one renewal, drifts to a script with one automation, and drifts to future you with one postponed decision — none of them worth a meeting on the day they happened. The tour exists so the drift gets noticed in daylight, while it is still cheap to argue with.
We like lists for the same reason we like lakeside evenings — they make a scattered thing walkable. If this one sent you thinking, our Bhopal travel guide does the same service for the city’s forts and lakes, our piece on planning a first trip to Bhopal turns wandering into an itinerary, and this short read on why Bhopal is called the city of lakes explains what a name gets right about a place. The work always lands somewhere. Better you chose the address.
