At my current job, we use in-line UPS units to power our servers. There is a couple advantages to this but the main ones include:
The servers get filtered power to them at a constant rate. Spikes in power don’t effect our servers because they plug into the batteries which are then plugged into the wall. In event of a brief power surge, they will keep going like nothing ever happened.
Main advantage is complete power outtage. Our servers have run times of over a couple hours each. Not just servers, but our networking equipment rack as well.
We have 3 networking racks with sets of 3 “SMX2000RMLV2U APC Smart-UPS X 2000VA Rack/Tower LCD 100-127V” units on each. The way it works is you place each battery on bottom and then the unit with the LCD display on top. You then plug battery 3 into battery 2 and then battery 2 into the controler unit.
In the event the batteries go bad in these, you just buy replacement batteries and hook them up as follows: