RTO and RPO: asking the right business questions before the architecture
RTO and RPO are not technical settings: they are business decisions that commit a budget. Asking them in the right order avoids paying for resilience nobody asked for — or missing it where it matters.
RGAIX team · · 7 min read
Two objectives that measure different things
The RTO (Recovery Time Objective) expresses a target duration to recover after an outage. The RPO (Recovery Point Objective) expresses the maximum acceptable data loss, measured in time. You can have a short RTO and a long RPO, or the opposite: they are two independent dials.
| RTO | RPO | |
|---|---|---|
| The question asked | How fast must we reopen? | How much work can we redo? |
| What lowers it | Standby environment, automation | Backup frequency, replication |
| What it costs | Capacity on standby | Volume and retention |
These are objectives to define and then to test. Typing them into a tool does not prove your system can meet them: until a drill has taken place, they are intentions.
Reason service by service
Order taking, a document intranet and a reporting tool do not share the same criticality. Applying a single RTO to the whole system means overpaying for half the scope and under-protecting the other half.
For each service, ask the teams three questions: how long an outage is genuinely bearable, what the operational consequences are beyond that, and which data would have to be re-entered. The third question reveals the real RPO — often far shorter than the infrastructure assumes.
Map the dependencies before promising anything
An application depends on a database, an identity system, the network, sometimes on third-party services. The recovery plan must include those dependencies and a restart order: rebooting a server does not restore a business journey.
- List the components the service depends on, external providers included.
- Identify the ones you do not control: their RTO becomes a floor for yours.
- Define the restart order and the checks between each step.
- Name the people required — and plan for their absence.
Test the restore, not just the backup
A drill checks the access, the backups, the procedures and the business behaviour after the restore. Record the observed time, the difficulties and the gaps against the objectives: that document is what makes the budget conversation concrete.
Surprises rarely come from the data itself. They come from expired access, a procedure written for an earlier version, a secret stored only on the machine that was lost, or a person who is unavailable.
Write it down, then review it
Have the objectives approved by both business and technical owners. State the scope, the assumptions, the responsibilities and how often they will be reviewed. A change of architecture, volume or supplier can invalidate an RTO that held last year.
A recovery objective only has value next to the means actually funded: without that, it is a reassuring line in a document nobody will reread.
Related pages
Maintenance, backups & disaster recovery
Prepare the continuity of your business.
Discover the service ↗READ NEXTProduction incident: what to prepare before you call
A short brief can save an hour on an incident. Here is what it contains — and what it must never contain.
Read the article ↗READ NEXTPreparing a rebuild: the SEO decision checklist
Most traffic losses after a rebuild come from decisions taken too late. Here are the ones to settle during framing.
Read the article ↗Let’s talk about what is blocking you.
An emergency, a profile to find or a project to structure: let’s start with your need.
