UPS for SMEs: 15 minutes of power that save the whole day

· 7 min read · SilverData team

Nobody thinks about the UPS until the power goes out at 11:40 on a Wednesday, right before the shipping cutoff. That's when it turns out the ERP server died without warning, the database is inconsistent, and the warehouse crew is standing around because the scanners have nothing to connect to. Four hours later the IT guy walks in with last night's backup and a question: what happened today, because it all needs to be re-entered.

The UPS is one of those infrastructure pieces that's boring, cheap compared to everything else, and skipped in every IT budget. And at the same time it's often the single point between "brief power blip, nothing happened" and "we lost half a day and the database is corrupted". We see both scenarios in practice — and the difference comes from decisions that take two hours and cost 3,000–8,000 PLN.

What a UPS should do (and what it won't)

An uninterruptible power supply isn't meant to keep the company running through an extended outage. That's what a generator is for, and it's a completely different price bracket. A UPS has three jobs: ride through short spikes and micro-outages (by far the most common event), give you time for a controlled shutdown of the server and network during a longer outage, and filter out dirty voltage from the grid, which slowly kills PC power supplies over the years.

Keeping the whole team working for an hour is not a realistic goal for a UPS in an SME. Keeping the server and the switch alive for 10–15 minutes so you can commit transactions to the database and shut the system down properly — absolutely. That difference in purpose drives every decision that follows.

Three UPS types — which one for what

The market splits into three classes that are easy to confuse in a shop, because they look similar and cost three times as much between the bottom and the top.

The second parameter — capacity, in VA or W. A sound rule: add up the draw of every device connected to the UPS and add a 30% margin. Server 400 W + switch 40 W + router 20 W = 460 W, so a UPS around 1,000 VA / 600 W. Buying "just in case" doesn't stretch runtime proportionally to price — better to match the load and, if needed, add an extra battery pack later.

Case study: what happens when there's no UPS

At one of our clients — a 30-person plant making automotive wiring harnesses — a distribution board on the first floor gave out mid-day. The server running Comarch Optima with a SQL database, sitting in a small room next to the office, died instantly, with no session close. When power came back twenty minutes later, SQL wouldn't start — the transaction log file was corrupted mid-write.

Recovery took 8 hours of work from two people: restoring from the overnight backup (SQL around 40 GB), re-entering 47 warehouse documents and 12 invoices issued in the morning, verifying stock levels. On top of that came one day of production where the warehouse ran on paper because the scanners had nowhere to write. Total real cost after adding downtime and staff time: about 14,000 PLN.

After the incident we installed a line-interactive 1500 VA UPS in the server room for the server, switch and router (3,200 PLN), configured automatic Windows Server shutdown after 8 minutes on battery, added a second 800 VA UPS for the VoIP PBX and the warehouse access point (900 PLN), plus the assessment and documentation (1,800 PLN). Total: 5,900 PLN — less than that one incident. In 14 months since — three brief power blips, zero data corruption, zero downtime.

What exactly to protect with a UPS

A common mistake: the bought UPS ends up powering a printer and a monitor, while the server sits directly in a wall socket. The priorities in an SME should look the opposite way.

  1. Server(s) — always first, together with the disk array. Even if the database is an Access file on a share, a corrupted file after an abrupt shutdown can block work for hours.
  2. Network switch and router — without the network, the server serves no one anyway. The internet router is often the forgotten device, and without it VoIP, e-commerce and integrations with the bank or e-invoicing all stop.
  3. VoIP PBX and Wi-Fi access points if the business depends on them for customer service or warehouse operations.
  4. Selected workstations — reception, POS terminals, the production line PC, the station that issues shipping documents.

Ordinary office workstations usually don't need a UPS — laptops have built-in batteries, and the accountant's desktop can tolerate a 20-second power blip without consequences if the file is saved. What's worth doing instead is fitting every workstation with a surge-protected power strip — an investment of 60–120 PLN that saves PSUs from slow death by voltage spikes.

Maintenance: batteries and testing

This is the part that's not on the price tag. A UPS battery ages and typically needs replacement after 3–4 years, whether it was used or not. After five years a UPS on the original battery gives you a real 60 seconds of runtime instead of the advertised 15 minutes — and nobody notices until the power goes out.

The practical routine we recommend to clients:

What next

The UPS is one of the best cost-to-protection ratios in an SME IT budget. It won't replace backups or a business continuity plan, but it fills a gap nothing else fills — the ten minutes of calm in which the server gets a chance to breathe out.

If you're not sure what's sitting in your server room today without a UPS, or what's connected to a six-year-old unit still running its original battery — start with an inventory. An hour with a multimeter and a list is enough to see where the largest risk lives.

A UPS won't keep the business running through a one-hour outage. It should keep the server alive long enough to shut down politely — and that's enough to save the day.

← Back to blog