Sysadmin Archetypes: What They Really Cost Your Company

Diagnosis: why “stable” on a resume should worry you

Monday, 9:40 AM. HR sends you a new sysadmin’s resume. Degree, certifications, three jobs, everywhere “everything ran stable.” You’re pleased. Three months later it turns out “stable” means nobody touches the server, because the admin himself is afraid to touch it.

Sysadmin archetypes aren’t a classification by skill level, like most articles claim. They’re about how a person behaves when everything is on fire. That behavior is what actually determines a specialist’s real cost to the business, not the salary written in the contract.

Sysadmins are medicine. Good medicine heals your infrastructure. But like any medicine, the side effects don’t show up in the instructions, they show up in your P&L. Over 25 years in this industry, I’ve hired, fired, and cleaned up after pretty much every archetype below. None of them are made up. Each has a real prototype whose name I obviously won’t share.

Below are five pure archetypes plus one bonus. For each one: the core trait, the approach to work, the moment your eye starts twitching, a real-life dialogue, and the outcome that costs the company money. At the end: a table with hidden costs and my own personal failure trying to combine archetypes by hand.

1. The Conservative Dinosaur: legacy is sacred

The core trait. 20 years of experience, watched Windows 95 get set up, and has genuinely believed progress stopped right there ever since. Cloud is hype, virtualization is a fad, Kubernetes is the thing that will break his RAID array.

The approach. Any problem gets solved by turning it off and on again. Didn’t help? Reinstall, preferably from the discs he keeps in a safe next to his employment record book. New software doesn’t get installed, on principle: “the old version is more stable.”

The trigger. Management decides to move to Linux or rent a VM in the cloud to stop paying for electricity in a server room bigger than the meeting room.

The outcome. They hire a junior to help. The Dinosaur sabotages: “these devops guys will break my RAID.” A month later the junior quits, and his machine gets a caring fresh install of Windows Server 2003. “See, now everything’s in order.”

The Dinosaur isn’t a villain. He’s just stuck in the moment when legacy really was sacred, because there was no alternative. The problem is, that moment ended about fifteen years ago, and you’re paying his salary for today.

2. The Backup Paranoid: keeper of three copies

The core trait. Doesn’t sleep at night. Convinced that tomorrow the server room will burn down, the RAID controller will die, and the gravitational constant will suddenly change, and then everything will definitely leak through the disks.

The approach. Backs up literally everything: the break room kettle’s logs, temp files, the accountant’s browser cache. RAID 10 plus tape duplication, which he personally drives to a bank safety deposit box at his own expense, because “it’s safer that way.”

The trigger. A user lost a file off their desktop. Simple request: “get it back from the recycle bin.”

The outcome. The file turns up an hour later. The user never writes to this admin directly again, only through a ticket. But now the Paranoid has an archived conversation proving he warned everyone.

Of all the archetypes, the Paranoid is the cheapest in the long run. The company doesn’t pay for his nerves, but it absolutely pays for a real data loss incident, and once is usually enough to remember the lesson. If you don’t have a single person like this on your team, read about how I lost part of an archive myself after deciding RAID counts as a backup. A cheap lesson that turned out not so cheap.

3. The Lazy Scripter: one click solves everything

The core trait. IQ 140, hates manual labor. Eight hours writing a script for a task that could be done by hand in five minutes isn’t wasted time to him, it’s an investment in dignity.

The approach. Any request gets “I’ll write a script.” The desktop is a console, the mouse is never used on principle, it “slows down thinking.”

The trigger. Accounting brings a list of 300 employees and asks for access to a new folder. By hand: thirty minutes of work.

The outcome. The script is ready, but by mistake it grants permissions not to the folder, but to the entire root of drive D:. The system goes down. To be fair, ten minutes later he rolls back the changes with the same script, because he’d thought to enable logging. The closing comment on the ticket: “Automated. Further inquiries not possible, the script also removed permission to create tickets.”

Funny right up until you count the downtime cost. Automation without granting root access to the whole drive is done differently, and I have a collection of PowerShell scripts for everyday network engineering tasks.

4. The Pessimist Philosopher: hardware karma

The core trait. An IT buddhist. He’s made peace with infrastructure being pain, and treats every outage as karmic payback. Never panics, just delivers the verdict in a quiet voice.

The approach. Asked “when will it be fixed,” he answers “whenever the cards fall that way.” Instead of a network diagram he draws something resembling a mandala made of patch cords, and that, by the way, is the only documentation the company has.

The trigger. A citywide power outage or the main router going down, the exact moment when documentation was needed yesterday.

The outcome. Three hours later the site comes back on its own, the provider just fixed the backbone line. “See? That was the universe testing us.” The manager leaves with a nervous tic, and the budget for a second power supply never materializes, until the next outage.

The only thing that actually saves you from hardware karma isn’t meditation, it’s monitoring that tells you about a problem 20 minutes before the client sees it. I’ve got a guide to setting up Uptime Kuma without extra hassle, it goes up in an evening, and the stars don’t need to align for it.

5. The Master of Regulations: IT’s bureaucrat

The core trait. Legally bulletproof armor against idiots. Never does anything “verbally,” only by written request, preferably in triplicate.

The approach. A request to replace a mouse takes him five pages, including a consent form for processing personal data via that very mouse.

The trigger. An emergency situation: a server light is on fire, the director is yelling “cut the power.”

The outcome. The server burns down. The Bureaucrat writes a memo stating the director failed to submit the request on time. He doesn’t get fired, on paper he’s completely in the right.

There’s a rational kernel in this archetype worth half the business: when IT is outsourced, or a contractor is used to working “on trust,” this kind of bureaucracy is the only thing protecting the company when something goes wrong. The other question is where the line falls between procedure and an actual fire in the server room. I covered this with examples of remote access and contractors who disappeared along with the passwords.

Bonus: the Combat Handyman, Steve-with-a-ladder

The core trait. Has no clue about networks on principle, but knows how to talk to the accounting department in their language and carries heavy monitors up the stairs without an elevator.

The distinguishing feature. Solves every problem through physical presence. If Steve showed up, people are already bringing him tea and cookies, and the problem seems almost embarrassed to happen again.

The computer turns on. The client is delighted, and the explanation obviously has nothing to do with the physics involved.

The outcome. He never gets promoted to lead engineer, because without Steve the office would freeze solid on day one of his vacation. He’s not an archetype of a problem. He’s the archetype of businesses sometimes needing the guy who shows up more than the guy who’s right.

Diagram: which archetype covers which risk

Before you get to the numbers in the table, look at the selection logic. It’s simpler than it sounds: don’t look for the perfect sysadmin, look for the one that fits your specific risk.

What this actually costs the business

These aren’t numbers invented for drama, they’re an average picture across companies with 50-150 employees that I’ve dealt with one way or another over the past few years. Currency is rubles, period is a year.

The exact numbers are approximate, but the order holds: the most expensive person isn’t the slow one, it’s whoever’s archetype blocks changes that were overdue a long time ago. The Dinosaur is the most financially ruinous character here, even though he looks the most harmless.

My own failure with this cast of characters

I once put both a Dinosaur and a Lazy Scripter on the same team, thinking the first one’s conservatism would balance out the second one’s risks. The logic seemed simple: one slows down changes, the other speeds them up, and you get a golden middle ground.

What I got instead was a four-month war between legacy batch-file scripts and PowerShell automation. Updating one broke the hacks in the other, and I spent that time untangling two parallel incidents instead of shipping one normal release. By my estimate, it cost around 400,000 ₽ in team work hours. And that’s not counting the gray hairs.

The conclusion is simple and unpleasant: different archetypes don’t automatically cancel out each other’s weaknesses. You have to deliberately combine them for a specific task, not hope opposites will balance each other out on their own. That logic doesn’t work with sysadmins, you need a task tracker, not dialectics.

FAQ

How do you tell if a company has a bad sysadmin?

Don’t look at diplomas, look at behavior under pressure: if every request turns into an argument about procedure instead of a solution, you’re not necessarily looking at a “bad” sysadmin, more likely one whose archetype doesn’t fit your risk profile. The second red flag: no documentation anyone besides him could understand.

How much does keeping one sysadmin cost a company of 50-150 people?

The salary is only part of the total. As the table above shows, hidden costs from a mismatched archetype can add 100,000 to 800,000 ₽ a year on top of the paycheck, through downtime, legacy license overpayment, or emergency call-outs.

Do you need an in-house sysadmin if IT is outsourced?

Yes, if incident response speed matters to you. Outsourcing almost always works through procedures and tickets, that’s the Bureaucrat archetype by default. It removes legal risk, but won’t save you in the moment when every minute counts.

What do you do if a sysadmin is sabotaging changes?

First figure out whether it’s a Dinosaur acting out of fear of anything new, or reasonable caution because of a real risk to legacy systems. Next, you need a concrete migration plan with dates and a rollback option, not a vague “let’s move forward.” A Dinosaur will resist a vague plan forever.

What’s the difference between a bad sysadmin and a mismatched archetype?

A bad one can’t do the job at all. A mismatched one can do the job fine, but his strengths don’t align with your company’s actual risks, a Paranoid is great where data is critical and overkill where there’s barely anything to lose.

Can you keep two archetypes on staff for balance?

You can, but don’t expect them to balance each other out automatically. Without a clear split of responsibilities, you get a war over architecture instead of synergy, I’ve been through that one, and it wasn’t cheap.

Outlook

None of these archetypes is bad on its own. It’s bad when it’s the only one in the company and there’s no plan for a backup. The Dinosaur is needed for legacy systems that genuinely shouldn’t be touched. The Paranoid belongs where a data mistake costs more than the whole team’s salary. The Lazy Scripter is invaluable for routine work that repeats dozens of times a month. The Philosopher is good because he doesn’t throw a fit during an outage, he just needs someone nearby with monitoring and a procedure. The Bureaucrat saves you when dealing with contractors. And Steve-with-a-ladder often holds an entire small business together, something that business only finds out the day Steve goes on vacation.

If I were building a team from scratch right now, I wouldn’t look for the “perfect sysadmin,” there’s no such thing, just different sets of side effects. I’d honestly ask myself: which archetype do I need for my company’s specific risk right now, not which one sounds more impressive on a resume.