Skip to main content
Azure Resiliency Map
Service catalog
Networking

Load Balancer / Application Gateway / Front Door

Load Balancer and Application Gateway are regional services — zone resiliency only. Azure Front Door is the actual cross-region failover mechanism for HTTP(S) traffic, via automated health-probe-based routing.

SLA 99.99% (Standard Load Balancer, zone redundant) / 99.95% (Application Gateway v2) / 99.99% (Front Door)Last verified 2026-08-22
Local

Basic Load Balancer (retired) / non-zonal frontend

Basic Load Balancer was retired on 30 September 2025 — existing deployments keep running but are unsupported and carry no SLA. A non-zonal Standard LB or Application Gateway frontend is still pinned to one datacenter.

Replication
None
RPO
N/A — routing layer, not a data store
RTO
N/A if the datacenter fails — no zone protection
Failover trigger
N/A
Relative cost
No extra cost

Baseline — a non-zonal Standard LB/App Gateway frontend costs the same as a zone-redundant one; the retired Basic SKU cost even less but is no longer a real option.

If you find a Basic SKU Load Balancer in production today, it is past end-of-support and should be migrated to Standard as a priority finding.

Meets tier
T0
T1
T2
T3
T4

Shows Unknown at every tier — there's no fixed RPO/RTO for this option; it depends entirely on an architecture you'd have to design and build.

Zonal

Standard Load Balancer / Application Gateway v2 (zone redundant)

Zone-redundant frontend IP configuration spreads the service across zones within the region.

Replication
None
RPO
N/A — routing layer
RTO
Seconds — automatic
Failover trigger
Automatic
Relative cost
No extra cost
Protects against: Datacenter/zone failure

Zone-redundant frontend IP is a configuration choice on the same Standard SKU — not a pricier tier.

Both are single-region resources — neither has a native concept of failing over to another region.

Meets tier
T0
T1
T2
T3
T4
Regional

Azure Front Door (priority/latency routing + health probes)

Global anycast service. All Front Door configurations include automated backend health monitoring and global failover — set backend priority and Front Door reroutes traffic away from an unhealthy region automatically based on health probe results.

Replication
N/A
RPO
N/A — routing layer only; the RPO for your application still depends entirely on the data tier behind each backend
RTO
Probe-interval driven — typically well under a minute once a backend is marked unhealthy
Failover trigger
Automatic
Relative cost
Moderate cost premium
Protects against: Region-wide backend outage, for routing purposes only

A separate billed service (base fee plus data processed/routing rules) on top of whatever backends it fronts — real money, but far cheaper than duplicating compute.

Front Door solves the traffic-steering half of regional DR. It does nothing for data consistency between regions — that's still on the services behind it.

Meets tier
T0
T1
T2
T3
T4

Gotchas

high

Basic Load Balancer is already retired

Retired 30 September 2025. It still technically runs if you have one, but it is unsupported and explicitly outside any SLA — anyone relying on its uptime commitment as part of a Tier assessment has no commitment to point to.

medium

Load Balancer and Application Gateway are regional, full stop

Neither has any concept of a secondary region. True multi-region HTTP(S) resilience requires Front Door (or Traffic Manager for non-HTTP/lower-layer scenarios) in front of independent regional deployments.

medium

Front Door failover is routing only, not a data guarantee

An automatic, fast Front Door failover to a healthy region can still serve stale or inconsistent data if the backend's own data replication hasn't caught up — routing speed and data RPO are two separate problems that get conflated in architecture reviews.

high

Neither Front Door nor Traffic Manager can health-probe an internal-only endpoint

Microsoft's own docs state it plainly: Traffic Manager "isn't designed to probe endpoints that resolve to addresses within private, non-routable, or Microsoft-internal network spaces," and such endpoints "must be configured as Always serve traffic endpoints" — meaning no health-based failover at all, just static routing. Front Door Premium can reach a private origin via Private Link, but its own entry point is still public. There is no first-party managed GSLB for a service that must never be internet-reachable — see the GSLB guide for what teams build instead.

Sources