Two different limits

Recovery time objective (RTO) is the maximum acceptable delay before a service is restored. Recovery point objective (RPO) is the maximum acceptable amount of data loss, usually expressed as the time since the latest usable recovery point. If an order system has a four-hour RTO and a 15-minute RPO, the plan must restore service within four hours and recover records to within 15 minutes of the interruption. These are decisions about business impact, not settings that a backup product can choose correctly on its own.

Different systems can need different targets. A public catalogue may tolerate a longer outage than payroll or an order queue. A workload may also depend on other services whose recovery times constrain its own. AWS’s guidance recommends setting objectives from business impact and warns against arbitrary or unrealistic targets, including assuming that zero downtime and zero data loss are achievable without checking the design.

Measure the restore, not the job status

A green backup status confirms that a job reported success; it does not show that the data is complete, readable or quick to restore. AWS recommends periodic recovery tests in a separate environment, checking that restored data is accessible and uncorrupted, comparing its newest records with the RPO, and measuring elapsed recovery against the RTO. A file-level spot check may be enough for a small personal archive; a production service may need a runbook that restores databases, configuration, identity access and application code in order.

For a practical exercise, choose one representative system, note the timestamp of its latest recovery point, restore it to an isolated target, and ask a user or automated check to retrieve real data. Record the time from incident declaration to usable service, not just the cloud job’s completion time. If the restore misses the target, identify whether the gap comes from backup frequency, network transfer, missing dependencies, credentials or manual steps. Change either the architecture or the agreed objective, then repeat the test. Recovery confidence comes from demonstrated recovery capability tied to a stated impact tolerance.

Sources & further reading

AWS recovery objectives AWS periodic recovery tests

Published by Capvise Media. Editorial standards · Corrections