From a distance, a server looks like any other plug-in appliance. It has a power lead, a plug and a casing. So why can it not just be tested along with the rest of the office?

The short answer: some testing requires equipment to be switched off and disconnected, and switching off business-critical equipment has consequences well beyond the box itself.

Testing can mean downtime

Many inspection and test steps need the item to be isolated from supply. For a printer, that is a minor inconvenience. For a server, a storage system or a core network switch, it can mean an outage for everyone who depends on it.

Live servers and UPS-supported systems cannot always be fully tested while operating. Where the approved method or access is unsuitable, the right outcome is to defer the item and record why, not to mark it as passed.

Redundancy is not a free pass

Dual power supplies and redundant feeds are designed to keep equipment running if one path fails. They do not automatically make it safe to disconnect one supply for testing:

  • Both supplies may not be connected to genuinely independent circuits.
  • A second supply may be faulty, overloaded or misconfigured without anyone knowing.
  • Other equipment may depend on the same PDU or UPS.

Assuming redundancy works, and finding out during a test that it does not, is exactly the kind of surprise a change process is meant to prevent.

Small devices can be critical too

It is not only the big rack equipment. A small switch under a desk might connect the reception phones. A compact firewall might be the only way the office reaches the internet. Do not assume a device is safe to unplug because it looks like ordinary office equipment.

What good planning looks like

Before any intrusive work on critical equipment, the following should be agreed in writing with your IT lead:

  1. Equipment scope. Exactly which assets are included, and which are not.
  2. Manufacturer guidance. Some equipment has specific instructions about testing methods.
  3. Power dependencies. Which services are affected when each item is disconnected.
  4. Shutdown requirements. Who shuts systems down, in what order, and how long it takes.
  5. Permitted methods. What testing is allowed on each item.
  6. Change approver and IT lead. Named people who approve the work and are available during it.
  7. The maintenance window. When work starts and when it must finish.
  8. Rollback and stop conditions. What happens if something does not go to plan, and when work stops.
  9. Restoration checks. Who confirms systems and services are healthy afterwards.

Your IT team remains responsible for backups, application shutdown, recovery and confirming service health. If the window runs out or approval is not in place, intrusive work stops; it is extended only with written authorisation.

UPS units need individual assessment

UPS systems and rack PDUs vary widely in configuration and manufacturer guidance. They need to be assessed individually. Internal UPS servicing and battery runtime, capacity or load-bank testing are separate specialist services and are not part of routine electrical testing.

What the records should show

A good infrastructure test record makes the limits clear. It shows what was tested and the result, what was visually inspected only, what was deferred and why, and what was excluded. It references the approved change. A visual-only inspection is never presented as a completed electrical test.

In short

  • Critical equipment testing can require downtime.
  • Redundant power does not automatically make disconnection safe.
  • Scope, approvals and a maintenance window come first.
  • Items that cannot be tested safely are deferred, not passed.