Active-Active (highest):Deploy to two active regions and distribute traffic between them.Active-Passive (medium):Deploy to main region and keep DR region on-standby. Upon disaster failover to secondary region.Active, Backup and Restore (low):Deploy to main region and backup required settings and data. Upon disaster create all the requiresed services in a secondary region and restore the settings and data.
- Don't forget about SLAs for infrastructure componets (VNEts, Application Gateways, Firewall, etc.).
- What about dependecies to on-prem systems?.
- Offered SLA may not improved with multi-zone deployment and does not guard against a region wide outage.
- Resiliency design patterns
- Application resiliency design patterns
- Circuit breaker, retry back-off, etc.
- Deploy your environments using IaC
- Leverage technologies such as ARM, Bicep and Terraform
- Don't forget your governance and security settings (RBAC, NSGs, diagnostic settings, log analytics workspaces)
- If your architecture depends on services that may take a long time to install, determine if you need to keep these services in a passive state
- Some services include:
- API Management
- Large AKS clusters
- Application Gateway
- VPN Gateway
- Deploy your workloads with DevOps pipelines
- Leverage data replication
- Review, plan, and test your failovers
- Perform a retrospective and incorporate learning from the the rehersal