Skip to content

Resilience Targets: RTO and RPO

Use this page to set recovery expectations before choosing topology, replication, and failover patterns.

Suggested targets by workload tier

Workload tier Typical business meaning Target RTO Target RPO Recommended posture
Critical Revenue, safety, or highly regulated workflows Minutes to less than 1 hour Near-zero to minutes Multi-region design, tested failover, strong automation
Important Core internal or customer-facing capability Less than 4 hours Minutes to 1 hour Zone redundancy plus defined recovery plan
Standard Business-supporting but tolerable interruption Same day Hours Single-region resilience with backup and restore drills
Dev/Test Non-production experimentation 1 to 2 days 1 day or more Minimal recovery investment

Azure service SLA orientation

Service type Reliability note
Zonal or zone-redundant PaaS Better resilience to datacenter-level faults where supported. [Documented]
Regional PaaS without multi-region failover Recovery is limited by regional outage posture and service capabilities. [Documented]
Multi-region active-passive Lower complexity than active-active but slower failover. [Inferred]
Multi-region active-active Best steady-state availability posture, highest design and ops complexity. [Correlated]

Single-region vs multi-region RTO comparison

Pattern Expected recovery profile
Single-region with backups Longer RTO and higher data-loss risk under regional outage. [Inferred]
Single-region with zone redundancy Better intra-region fault tolerance, little help for region loss. [Documented]
Multi-region active-passive Faster recovery if failover is rehearsed and dependencies are paired. [Validated]
Multi-region active-active Fastest customer recovery when data and routing patterns support it. [Assumed]
flowchart LR
    A[Business tier] --> B[Set RTO and RPO]
    B --> C{Regional outage tolerance?}
    C -->|Low| D[Single region or zonal]
    C -->|Very low| E[Multi-region]
    D --> F[Backup and restore drills]
    E --> G[Failover automation and exercises]

Microsoft Learn references

  • https://learn.microsoft.com/en-us/azure/reliability/overview
  • https://learn.microsoft.com/en-us/azure/well-architected/reliability/