Begin with the services people need
A power cut can affect more than a server. The network, firewall, internet connection, phones and access to cloud services may all depend on equipment at your site. The useful starting point is to identify what must keep working briefly, what should shut down safely and how people will work if the site stays unavailable.
List the systems and connections that support essential work. Include onsite servers and storage, network equipment, connectivity, telephony and any local equipment needed to reach cloud applications. Decide who can declare an outage, who speaks to staff and suppliers, and which tasks can continue another way. Those decisions determine what needs temporary power and what needs a recovery plan.
Check the UPS and the shutdown sequence
An uninterruptible power supply (UPS) can bridge a short interruption and give equipment time to shut down in a controlled way. It is not a promise of a fixed number of minutes. Useful runtime depends on the connected load, battery condition and system design. The shutdown software and any virtual-machine sequence must be configured and tested for the actual equipment.
Schneider Electric's UPS management guide describes how compatible software can initiate graceful shutdown and, in some configurations, sequence virtual machines. Merely plugging in a UPS does not establish that these steps will happen correctly.
For schools and colleges, the DfE servers and storage standard calls for critical servers to have automatic safe shutdown and at least 30 minutes of UPS runtime. That is a school-specific standard, not a general runtime rule for every business.
Do not confuse resilient disks with a backup
Redundant disks can help a server continue after a disk fails. They are not a separate recoverable copy if data is deleted, corrupted or encrypted by an attacker, and they do not make the whole site immune to a power cut. Redundancy, replication and backup address different failures. Microsoft's reliability guidance explains the distinction.
Ask what would happen if a server, storage device or building were unavailable. A design that tolerates one failed disk may still depend on a single power supply, network connection or site. The recovery plan should reflect the failures that matter to your organisation.
Plan for an outage that lasts longer
Keep recoverable copies away from the same failure point, agree which services should come back first and test a representative restore. Know where the credentials, documentation and supplier contacts are held when the main site is down. The NCSC's guidance on backing up critical data explains why important information needs a separate recovery route.
A recovery plan also needs a communication route. Staff should know whom to contact, whether they can work from another location and what to do if a local connection prevents access to otherwise available cloud systems. The time and cost of a more resilient design depend on the systems, site and acceptable downtime.
Five questions for your IT provider
- Which critical equipment depends on power at this site?
- How much UPS runtime has been measured under the current load and battery condition?
- Will servers and other critical systems shut down safely, in the right order?
- Which services can staff use from another location during a longer outage?
- When was a restore last tested, and who owns the recovery steps?
If any answer is unclear, review the dependencies before an outage turns them into an emergency. CompTech can help discuss the scope of backup and recovery work for your environment; this guide does not promise a fixed amount of uptime or an unagreed service. Explore our backup and disaster recovery approach, or talk to us about a practical review.
Sources reviewed 28 September 2026: DfE servers and storage standard, Schneider Electric UPS management guide, Microsoft reliability guidance and NCSC backup guidance. Check the current guidance and your own equipment before making a design decision.
About this guide
This guide replaces the earlier article at this address with freshly written practical guidance and current primary references.
Use this guidance alongside your organisation’s policies and agreed IT support arrangements. Product features and licence terms can change; check the linked supplier guidance before making a purchase or changing settings.