App Service
Zone redundancy requires Premium v2/v3+ and a minimum instance count. There is no native regional failover feature — multi-region App Service is Front Door/Traffic Manager plus a second, independently deployed app.
Standard/Premium tier, no zone redundancy
Multi-instance plan within a single zone. The platform auto-heals unhealthy instances but the whole plan sits in one datacenter.
- Replication
- None
- RPO
- N/A — app tier is stateless; state lives in whatever backing data store you use
- RTO
- Minutes — platform detects and replaces unhealthy instances automatically
- Failover trigger
- Automatic
Zone-redundant Premium v2/v3 (or Isolated v2)
Instances automatically spread across Availability Zones. Requires a minimum instance count (2–3 depending on tier).
- Replication
- None
- RPO
- 0 for the app tier itself
- RTO
- Seconds to minutes — automatic instance rerouting across zones
- Failover trigger
- Automatic
Only available on Premium v2/v3 and Isolated v2 — Standard tier has no zone-redundant option at all.
Front Door / Traffic Manager + duplicate deployment
No native geo-failover. The pattern is: deploy the app independently to a second region, then use Front Door (health-probe based) or Traffic Manager (DNS based) to route around a regional outage.
- Replication
- N/A
- RPO
- Entirely dependent on the backing data store(s) behind the app — App Service itself holds no state to lose
- RTO
- Routing failover: tens of seconds to a couple of minutes once configured — but keeping the second region's code, config, and secrets in sync is your CI/CD pipeline's job, not a platform feature
- Failover trigger
- Automatic
Gotchas
Zone redundancy has a minimum instance count you can miss
Enabling zone redundancy below the required instance count (2 for some tiers, 3 for others) either fails outright or silently doesn't distribute the way you expect — always verify actual instance placement, not just the toggle.
The SLA percentage doesn't prove zone coverage
99.95% is the headline figure for both zonal and non-zonal Premium plans — zone redundancy buys resiliency to a datacenter failure, not a different SLA number. Don't use the SLA % alone as audit evidence that zone redundancy is actually configured.
"Multi-region App Service" is not a feature you turn on
There's no App Service geo-failover button. Every multi-region App Service architecture is Front Door/Traffic Manager plus a fully separate deployment, with its own release pipeline, config, and secret replication story that you own end-to-end.