How to Check UPS Battery Runtime and Plan Replacement for a Home Lab

Your UPS says "online." The self-test passes. The management software shows a runtime estimate that looks reasonable.

None of that proves the battery can carry your rack through a real outage.

A UPS status light typically confirms one thing: the unit can transfer to battery. It does not confirm the battery can sustain your actual load for the time your servers need to shut down cleanly. Battery capacity fades gradually, so the failure mode is rarely a clean break. It's an outage that lasts 40 minutes when you planned for 15, and your NAS was still flushing writes when the battery gave out.

This is a maintenance workflow, not a shopping decision. By the end, you'll have one of three answers: keep and monitor, replace the battery, or retire the unit. The goal is a measured runtime compared against your actual shutdown window — not a status indicator.

Why a Status Light Is Not a Runtime Measurement

A UPS self-test and a runtime measurement answer different questions.

A self-test or "online" indicator typically verifies that the UPS can switch to battery power and that the battery holds some charge. It does not tell you how long that battery can sustain your specific load under your specific conditions. A battery at 60% of its original capacity may still pass a self-test while delivering less than half the runtime you need.

Manufacturer runtime figures are claims, not measurements of your battery. They're based on a specific load, a new battery, and controlled conditions. Your battery is older, your load is different, and your ambient temperature may not match the test environment. The spec sheet tells you what the unit was rated to do when new. It says nothing about what your aging battery will do today.

Battery degradation is gradual. A lead-acid battery that delivered 20 minutes at your load two years ago might deliver 12 minutes now and 8 minutes next year. You won't notice the decline until an outage exposes it — usually at the worst possible time.

Separate what you know from what you're assuming:

• Documented: The UPS model's official test procedure, battery chemistry, and safe-handling guidance.
• Inferred: Battery age, replacement history, and physical condition.
• Measured: Actual runtime under your representative load.

Only the third one answers the maintenance question.

Define the Runtime You Actually Need

Before you test anything, write down a target runtime in minutes. Without it, any test result is just a number.

Start by identifying the protected load. That's every device plugged into the UPS: servers, NAS, network switches, router, modem, and any PoE gear that needs to stay up. Not everything on that list needs the same runtime.

Decide what you're actually buying time for:

• Ride out short blips. If your goal is to survive brief flickers and brownouts without rebooting, you need seconds to a couple of minutes. Most healthy UPS batteries clear this easily.
• Graceful shutdown. If your goal is to let the OS, hypervisor, and storage flush writes and power down cleanly, you need enough time for the full shutdown sequence. That's typically several minutes, but it depends on your stack.

Estimate your shutdown window honestly. A single NAS with a simple shutdown script might need two minutes. A virtualization host running multiple VMs, a database, and a storage array might need ten or more. Add margin — shutdowns take longer under load or when something hangs.

One detail that trips people up: network gear often needs to outlast the servers. If your shutdown commands travel over the network, your switch and router must stay powered long enough to deliver them. If the switch dies first, the servers never receive the shutdown signal and you're relying on their internal timers — if they have them.

Write down your target. Something like: "15 minutes at current load to allow full graceful shutdown of NAS and hypervisor, with network gear powered for the full duration."

That number is what you'll test against.

Establish a Representative Load Before Testing

A runtime test is only meaningful if the load during the test resembles the load during a real outage.

Start by measuring the actual watt draw of your connected equipment. A plug-in power meter between the wall and the UPS gives you a real number. If you can't measure directly, add up the rated wattage of each device — but understand that rated wattage is usually a maximum, not typical draw. Your server idles lower than its PSU rating.

This is where the VA-versus-watts distinction matters. VA (volt-amperes) is apparent power; watts is real power. A UPS rated for 1500 VA might only deliver 900 watts. The gap is the power factor. If you size your expectations around the VA number, you'll overestimate runtime. The watt rating is the one that determines how long your battery lasts.

Decide whether to test at full production load or a documented representative subset. Full load is more accurate but riskier — you're pulling power from live systems. A representative subset (say, the NAS and switch but not the idle workstation) is safer but less precise. Either way, record what you tested.

Account for load that changes during an outage. Spinning disks draw more during spin-up. A CPU under load draws more than one at idle. PoE devices pull power continuously. If your load fluctuates, your runtime will too.

Record the load figure alongside your test result. When you re-test in six months, you need to know whether the runtime changed because the battery aged or because you added a device.

Inspect the Battery and Unit Before You Test

Before you run any test, look at the hardware. Some conditions are stop signs, not test conditions.

Check for:

• Swelling in the battery case
• Corrosion on terminals or connectors
• Scorch marks or discoloration
• Excessive heat when the unit is idle
• Burnt odor near the battery compartment

If you find any of these, stop. A damaged battery is a safety issue, not a runtime question. Don't test it, don't keep it in service, and follow the manufacturer's guidance for safe handling and disposal.

Read the battery manufacture date. It's usually printed on the battery label. Also check any replacement history — if you don't know when the battery was last replaced, assume it's original. Batteries self-discharge over time, and a battery left uncharged for extended periods can be permanently degraded.

Confirm the battery chemistry. Lead-acid (VRLA) and lithium-ion units differ in expected service life and in what testing is appropriate. Lithium-ion systems may have battery management electronics that change how you should test and replace them. Follow the specific manufacturer's guidance for your chemistry — there is no universal procedure.

Before deciding the unit is worth keeping, verify that a compatible replacement battery is actually available. If the battery is discontinued and no equivalent exists, the decision may already be "retire" regardless of how the runtime test turns out.

Run a Controlled Runtime Test Safely

There is no single universal runtime test procedure. Different UPS models and battery chemistries have different supported test methods, prerequisites, and warnings. Use your specific UPS model's manual for the procedure. What follows is a framework, not a substitute for the manufacturer's instructions.

Set your stop condition before you start

The most important preparation step is deciding in advance when the test ends. A home-lab runtime test is not a full-discharge experiment. Its purpose is to answer one question: does the battery carry your load for at least as long as your shutdown window requires?

Set your stop condition as your required shutdown window plus a safety margin. If your target is 15 minutes, plan to stop the test at 15 minutes of battery operation — or at 18 minutes if you want a small buffer against measurement error. Do not run the test to UPS cutoff with live equipment attached. A full discharge to cutoff is a different procedure that belongs in an isolated, protected test setup under the manufacturer's guidance, not on production gear.

Write the stop condition down before you unplug anything. It removes the temptation to "just see how long it goes" while your NAS is running on battery.

Prepare first

• Confirm the battery is charged according to the manufacturer's guidance. Some manufacturers specify a charge condition before runtime testing.
• Save or verify your shutdown configuration. If the test triggers a shutdown, you want it to be the one you configured, not a surprise.
• Decide how to protect live systems during the test. If you can't tolerate downtime, test with a representative subset or schedule the test during a maintenance window.
• Notify anyone who depends on the services you're about to put at risk.

Perform the test

Remove AC input from the UPS — unplug it from the wall or switch off the circuit. The UPS should transfer to battery. Start a timer. Observe:

• How long the UPS carries the load before you reach your stop condition
• Any alarms or warnings during the discharge
• Whether any devices shut down early or behaved unexpectedly
• Whether the UPS shed load (some units drop non-critical outlets first)

When you reach your stop condition, restore AC input. Do not wait for the UPS to shut down on its own. Record the observed runtime, the load at the time, the battery age, and the ambient conditions. These details make the result comparable next time.

Know what you're testing

A short transfer check — unplug, confirm the UPS switches to battery, plug back in — answers "does backup work at all?" It takes seconds and carries minimal risk.

A timed runtime test — let the UPS run on battery until you reach your pre-set stop condition — answers "does it last long enough?" It takes minutes and carries real risk to live systems, which is why the stop condition matters.

These are different tests. A passing transfer check does not prove your shutdown window is covered. And a timed test that stops at your required window plus margin gives you a clean yes-or-no answer without discharging the battery further than the question requires.

Do not run a full discharge on production equipment without a plan. A test that causes data loss or an unplanned outage defeats its purpose. If you can't safely test at full load, test what you can and document the limitation.

Interpret the Result Against Your Shutdown Window

Now compare your measured runtime to the target you defined earlier. Not to the manufacturer's spec sheet. Not to what the management software estimates. To your number.

• Comfortably above your shutdown window: The battery is healthy for now. Keep it and re-test on a schedule.
• Near or below your shutdown window: The battery is a replacement candidate. It may still pass a self-test, but it no longer provides the margin you need.
• Far below, or the UPS shut down almost immediately: The battery is degraded or failing. Replace it or retire the unit.

A single test is a snapshot. Battery capacity continues to decline, and the rate of decline can accelerate as the battery approaches end of life. One good result today doesn't guarantee a good result in six months. The trend matters more than any single reading.

Monitoring software can reduce disruption and help you track runtime estimates over time. But a status readout is not a substitute for a measured runtime. The software estimates based on the battery's reported state, which may not reflect actual capacity under load. Use monitoring to watch for changes and to trigger re-tests — not as the final answer.

One more caveat: a short test may confirm that backup works while still failing to prove the required shutdown window. If you only tested for two minutes and your shutdown takes eight, you haven't answered the question.

Replace the Battery or Retire the UPS

The decision boundary is straightforward once you have a measured runtime and a target.

Replace the battery when:

• The unit is otherwise sound (no damage, no scorch marks, no corrosion)
• A compatible replacement battery is available
• The measured runtime no longer meets your target
• The unit still suits the load it must protect

Retire the unit when:

• It shows physical damage or signs of internal failure
• It's unsupported by the manufacturer (no firmware updates, no documentation, no parts)
• No compatible replacement battery is available
• The unit's capacity or features no longer match your load
• The cost of replacement approaches the cost of a new unit with better capabilities

That last point deserves attention. A replacement battery can cost a meaningful fraction of a new UPS. When the unit is old, inefficient, or lacks features you now need — network management, pure sine wave output, better shutdown integration — putting money into a battery for a marginal unit may not be the best use of the budget. Compare the replacement cost against the value of the whole unit, not just against the inconvenience of shopping.

For lithium-ion systems, follow the manufacturer's safe handling and replacement guidance. Some systems require qualified service or have specific battery-management requirements. Don't improvise.

After replacement, re-run the same test under the same load. A new battery should restore the required runtime. If it doesn't, the problem may be the UPS itself — the charging circuit, the inverter, or the transfer switch — and the unit may need service or retirement.

Build a Repeatable Maintenance Routine

A one-time test tells you where you are. A routine tells you where you're heading.

Set a periodic check interval. For lead-acid batteries, annual testing is a reasonable starting point; some owners test more frequently as the battery ages. For lithium-ion, the manufacturer's guidance should inform the interval. Whatever you choose, put it on a calendar.

Log these fields each time:

• Date
• Measured runtime
• Load at time of test (watts)
• Battery age
• Ambient temperature
• Any observations (alarms, unexpected behavior)

The trend across entries is more useful than any single result. A runtime that drops from 18 minutes to 15 to 11 over three tests is telling you something a single reading can't.

Re-test after any change to the protected load. Adding a device shortens effective runtime. Removing one extends it. If you don't re-test, you're guessing.

Keep your shutdown configuration verified, not just your battery. Software updates, OS changes, and network reconfigurations can break shutdown integration without touching the UPS. Test the full chain — UPS triggers shutdown, shutdown completes cleanly — not just the battery.

Plan the replacement budget ahead of failure. Batteries don't fail on a convenient schedule. Knowing that a replacement is likely in the next year lets you budget for it rather than scrambling during an outage.

The routine is the deliverable. The goal is knowing your real runtime before you need it — not discovering it when the power goes out.

The Decision Rule

If measured runtime under your representative load clears your shutdown window with margin, keep the unit and re-test on a schedule. If it falls short but the unit is sound and a compatible battery exists, replace the battery and re-verify. If the unit is damaged, unsupported, or has no viable replacement, retire it.

The status light was never the answer. The measured runtime was.