What Is a UPS, and How Long Should One Actually Keep Your Business Running?

A UPS (uninterruptible power supply) is a battery between your equipment and the wall socket. When utility power fails, it carries the load for a few minutes so that servers, firewalls and phone systems shut down cleanly instead of dropping dead mid write. That is its entire job. It is not a generator, and it is not a backup plan.

Most small businesses buy one, plug it in, and never think about it again. That is where the trouble starts, because the battery inside is a consumable with a shelf life, and a UPS that has quietly failed looks exactly like a UPS that is working.

Think your IT is in good shape?

Take the free 3-minute readiness quiz

What does a UPS actually protect against?

Utility power fails in more ways than simply switching off. A UPS covers four of them:

• Complete outage.

  The obvious one. The battery takes over instantly and holds the load.

• Sag or brownout.

  Voltage drops below the level equipment expects, often when heavy machinery elsewhere in the building starts up. Equipment can reboot or corrupt data without the lights ever going out.

• Surge and spike.

  A short burst of excess voltage, typically from grid switching or a nearby lightning strike.

• Line noise and frequency variation.

  Electrical interference that shortens the life of power supplies over time.

A basic surge protector handles only the third of those, and only until its internal components wear out, which they do silently. A UPS handles all four and tells you it is doing so.

How long will a UPS keep a server running?

Less time than most people assume. Runtime is not a fixed property of the unit, it is a function of how much load you put on it, and the relationship is not linear.

The rule of thumb: a UPS sized so that your equipment draws roughly half its rated capacity will typically deliver somewhere between ten and twenty minutes. Load it to ninety percent of capacity and that can fall to four or five. Doubling the load more than halves the runtime.

That matters because ten minutes is not “ride out the outage” time. It is “shut down properly” time, and only if the shutdown is automated. Which leads to the part almost everyone gets wrong.

The mistake: a UPS with no shutdown configured

A UPS without automated shutdown is a delayed crash.

The unit carries the load for eight minutes, the battery exhausts, and everything drops at once with open files and mid flight database writes. The outcome is identical to having no UPS at all, except you paid for the box and replaced its battery twice.

Every business grade UPS can signal connected equipment over USB or network to begin a graceful shutdown when the battery reaches a set threshold. That requires software installed, configured, pointed at the right machines, and tested. In environments we take over, this is one of the most common findings: the hardware is present, the shutdown logic is not.

How long do UPS batteries last?

Sealed lead acid batteries, which is what sits inside most small business units, generally last three to five years. Heat is the main enemy. A UPS in a warm closet with poor airflow, which describes most small business server rooms, ages faster than the datasheet suggests.

The failure mode is the problem. A degraded battery still shows a green light and still reports “online” right up until the moment it is asked to carry load, at which point it delivers thirty seconds instead of fifteen minutes. The only reliable way to know is to test under load on a schedule and to replace on age rather than on symptoms.

Practical approach: record the install date on the unit itself, plan replacement at year three, and run a load test at least twice a year. If nobody owns that calendar, it will not happen.

UPS, generator, or cloud? They solve different problems

Protects against	UPS	Generator	Cloud hosting
Momentary sag or spike	Yes	No	Not applicable
Outage of minutes	Yes	Starts too slowly alone	Not applicable
Outage of hours or days	No	Yes	Yes, if staff can work elsewhere
Hardware failure or fire	No	No	Yes
Ransomware	No	No	Only with tested, isolated backups

Sites that need to stay running through a long outage pair the two: the UPS covers the gap while the generator starts and stabilizes. Sites that have moved workloads off local hardware need far less of both, because there is less on site left to keep alive.

Sizing a UPS without overthinking it

1. Add up the load.

   Total the wattage of everything that must stay up. Servers, the firewall, the core switch, the phone system. Not monitors, not printers, not space heaters.

2. Add fifty percent headroom.

   This buys runtime and leaves room for growth. Running a UPS near its ceiling is what produces four minute runtimes.

3. Decide what runtime you need.

   If shutdown is automated, ten minutes is usually plenty. If a generator is involved, you need to cover its start and transfer time, typically under a minute.

4. Check the outlet count and type.

   A common and avoidable problem is buying the right capacity in the wrong form factor for the rack.

Two smaller units split across separate circuits are often better than one large one, because a single UPS is itself a single point of failure.

Where a UPS fits in business continuity

A UPS buys minutes. Continuity planning is about what happens after those minutes run out, and that is a different exercise: which systems must come back first, how quickly, from where, and who decides.

The practical test is not whether you own a UPS. It is whether you can answer this: if the building lost power right now and did not get it back for two days, what would your team do tomorrow morning? Most businesses have never walked through that conversation, and the answer usually reveals dependencies nobody had mapped. The UPS is the smallest and cheapest piece of it, which is precisely why it gets bought and then forgotten.

Frequently asked questions

Do I need a UPS if everything is in the cloud?

Usually a smaller one, not none. Your internet connection, firewall, switches and access points are all still on site, and none of them work without power. If staff need to keep working in the building during a short outage, that equipment needs to stay alive.

Can I plug a laser printer into a UPS?

No. Laser printers draw very large current spikes when the fuser heats, which can overload the unit or trigger a fault. Use the surge only outlets if the unit has them, or a separate circuit.

What does the beeping mean?

Usually one of three things: it has switched to battery, the battery has failed a self test, or the unit is overloaded. All three are worth looking at the same day. Silencing the alarm without diagnosing it is how businesses discover the battery died eleven months ago.

How often should a UPS be tested?

Twice a year under real load, plus whenever equipment is added to the rack. A self test verifies the electronics, not the battery’s ability to carry your actual load.

Is a UPS a backup?

No, and conflating the two is a genuine risk. A UPS protects a running system for minutes. A backup protects your data when the system is gone. You need both, and the backup matters more.

Sources

Battery service life, runtime-to-load behavior and sizing headroom are general characteristics of sealed lead acid UPS equipment and should be confirmed against your specific unit’s datasheet, which will state its own runtime curve. Findings about missing shutdown configuration reflect what Corporate Technologies engineers encounter during onboarding assessments and are qualitative, not measured.

This article is general guidance on power protection and is not an electrical engineering assessment of any specific site. Sizing, circuit capacity and generator integration should be reviewed against your own equipment and local electrical code before purchase.