Business

Backup and Disaster Recovery for Business Applications: RPO and RTO in Plain Terms

November 11, 2026 · 5 min read · Etzal Group Team

Every company has backups until the day it needs one. Then it discovers that the job stopped in March, or that nobody knows the password to the archive, or that restoring takes three days. Backup and disaster recovery planning is the work of finding those things out on an ordinary Tuesday.

Two numbers that define the plan

RPO, recovery point objective. How much recent work can you afford to lose? If the answer is "one day of orders", a nightly backup is enough. If it is "nothing after a payment is confirmed", you need continuous replication or transaction logs.

RTO, recovery time objective. How long can the system be unavailable? A reporting tool can be down for a day. A point-of-sale system in a busy store cannot be down for an hour.

Set these per system, with the business owner of that system, in hours. They decide the design and the cost. Tighter numbers cost more, so it is worth being honest about which systems truly need them.

Classify your systems

A simple three-level scheme works:

  • Critical. Revenue or operations stop without it. RPO in minutes, RTO within a few hours.
  • Important. Work is slowed, and can continue for a day. RPO of a few hours, RTO within a day.
  • Supporting. Inconvenient to lose. RPO of a day, RTO of several days.

Write the list down. It tells you where to spend.

What to back up

People think of the database and forget the rest:

  • Databases, including transaction logs if your RPO is short.
  • Uploaded files such as documents, images and attachments.
  • Configuration: environment settings, scheduled jobs, certificates.
  • Secrets: keys and passwords, stored securely and separately.
  • Source code and build scripts, in a repository you control.
  • Infrastructure definitions, so servers can be rebuilt, not remembered.

A database restored without its uploaded files or its encryption keys is often useless.

The 3-2-1 rule

Keep three copies of your data, on two different kinds of storage, with one copy off site. In cloud terms: the live data, a backup in another storage service, and a copy in a different region or with a different provider.

Add two protections:

  • Immutability. At least one copy cannot be changed or deleted for a set period, even by an administrator. This is the defence against ransomware and against an attacker with stolen credentials.
  • Separate access. The account that runs the application should not be able to delete the backups.

Retention

Decide how far back you can go. A common pattern keeps daily backups for a month, weekly for a quarter and monthly for a year. Some problems, such as silent data corruption, are noticed weeks later, so a single recent copy is not enough.

Check retention against legal requirements, including how long records must be kept and how personal data deletion requests are handled in backups.

Test the restore

A backup is proven only by restoring it. Schedule tests and record:

  • Which backup was restored and how old it was.
  • How long the restore took from start to a usable system.
  • Whether users could log in and recent records were present.
  • What went wrong and what was changed afterwards.

Compare the measured time with the RTO. If the target is four hours and the test took nine, either the design or the target has to change.

Beyond backups: staying up

For critical systems, restoring from backup may be too slow. Options, in increasing cost:

  • Standby database replicating continuously, ready to take over.
  • Multiple servers behind a load balancer, so one failure is not an outage.
  • A second region that can serve traffic if the first is lost.

Replication is not a backup. A deletion or a corrupt update is replicated instantly. You need both.

The runbook

During an incident nobody thinks clearly. Write the steps in advance:

  1. Who declares a disaster and who is called.
  2. Where the backups are and how to reach them, including credentials stored somewhere that survives the outage.
  3. The restore steps in order.
  4. How to check the restored system before letting users in.
  5. How to redirect users and other systems to it.
  6. Who communicates with staff and customers, and what they say.

Keep a copy outside the system it describes.

Common gaps

  • Backup jobs that fail silently. Alert on failure and on a missing backup.
  • Backups stored on the same server as the data.
  • Encryption keys kept only inside the system being backed up.
  • No one person responsible.
  • A plan that depends on an employee who has since left.
  • Never having restored.

A reasonable start

For a mid-sized company with a handful of systems: classify them, set RPO and RTO for each, confirm that backups cover databases, files and configuration, put one copy out of reach of the application, turn on failure alerts, and schedule a restore test this quarter. That alone removes most of the risk.

Backup design, restore testing and runbooks are part of how we deliver and maintain systems in Enterprise Software. If you are unsure whether your current backups would hold, ask us for a review.

Summary

Set how much data you can lose and how long you can be down, per system. Back up everything needed to rebuild, keep an untouchable off-site copy, test restores and measure them, and write the steps down before you need them.

Frequently asked questions

What is the difference between RPO and RTO?

RPO, recovery point objective, is how much recent data you can afford to lose, measured in time. RTO, recovery time objective, is how long the system may be unavailable. A nightly backup gives an RPO of up to a day. Restoring it in four hours gives an RTO of four hours.

Are cloud services backed up automatically?

The provider protects its own infrastructure, which is different from protecting your data from your own mistakes. Accidental deletion, a faulty update or a compromised account can destroy data that the provider faithfully replicates. Check what your plan includes and keep an independent copy.

How often should we test a restore?

At least twice a year for critical systems, and after any significant change to the system or the backup method. A restore test should bring the system up from the backup in a separate environment and confirm that people can log in and that recent records are present.

Author

Etzal Group Team

Etzal Group Team, helping businesses grow with technology. Contact us.

Related service

Free consultation

Not sure where to start? We will map the right system for your business, free.