High availability and resiliency are among the most tested concepts in AZ-305. Many candidates confuse Availability Sets with Zones, DNS-based routing with edge routing, or backup with disaster recovery. This capsule highlights the common pitfalls, best practices, and exam strategies you must know.
Common Pitfalls
1. Confusing Availability Sets vs Zones
-
Availability Sets = rack-level redundancy.
-
Availability Zones = datacenter-level redundancy.
Exam Trap: If requirement is 99.99% SLA, answer = Availability Zones, not Sets.
2. Forgetting Region Pairs
-
Region Pairs are prioritized for recovery.
-
Some assume any two regions provide same benefits.
3. Misusing Traffic Manager vs Front Door
-
Traffic Manager = DNS-based, not real-time.
-
Front Door = edge routing, real-time failover.
Exam Trap: If scenario requires real-time failover, answer = Front Door, not Traffic Manager.
4. Storage Replication Confusion
-
LRS = local only.
-
ZRS = zone-level HA.
-
GRS = region replication but no read access.
-
RA-GRS = region replication with read access.
Exam Trap: If requirement is readable secondary in another region, answer = RA-GRS.
5. Database HA vs DR
-
HA (in-region) = Availability Zones or Always On replicas.
-
DR (cross-region) = Auto-Failover Groups or Geo-Replication.
Exam Trap: If requirement is cross-region automatic failover, answer = Auto-Failover Groups, not Geo-Replication.
6. Overlooking Application Resiliency Patterns
-
Many candidates forget retry, queue, circuit breaker patterns are required at the app layer.
-
HA infra alone won’t solve transient failures.
Best Practices
-
Always design across multiple fault/update domains (Sets) or multiple datacenters (Zones).
-
Deploy in region pairs for DR and compliance.
-
Use Traffic Manager for compliance routing, Front Door for real-time global failover.
-
Pick storage replication based on RPO/RTO needs.
-
Implement app-level resiliency patterns (retry, queues, circuit breakers).
-
Align with business SLA targets:
-
99.9% → Single instance.
-
99.95% → Availability Set.
-
99.99% → Availability Zones.
-
Quick Recall Table
| Requirement | Best Solution |
|---|---|
| Rack-level resiliency | Availability Set |
| Datacenter-level resiliency | Availability Zones |
| Region-level DR with compliance | Region Pairs |
| Real-time global failover & acceleration | Azure Front Door |
| DNS-based global routing | Traffic Manager |
| Read access to secondary storage region | RA-GRS |
| Automatic SQL failover across regions | Auto-Failover Groups |
| Global DB with tunable consistency | Cosmos DB |
| App-level retry on transient failures | Retry Logic + Circuit Breaker |
Exam Tips
-
“Which Azure feature provides 99.99% uptime for VMs?” → Availability Zones.
-
“Which storage option provides read access from another region?” → RA-GRS.
-
“Which service provides global routing based on DNS?” → Traffic Manager.
-
“Which service provides real-time failover at edge?” → Front Door.
-
“Which DB feature supports automatic cross-region failover?” → Auto-Failover Groups.
-
“Which resiliency pattern avoids hammering a failing service?” → Circuit Breaker.
What to Expect in the Exam
-
Direct Q: “Which feature protects against datacenter-level failures within a region?” → Availability Zones.
-
Scenario Q: “E-commerce app must survive full region outage with real-time failover.” → Azure Front Door + Region Pair.
-
Scenario Q: “Company needs SQL DB failover across two regions with single listener endpoint.” → Auto-Failover Group.
-
Scenario Q: “Company wants users to continue reading product catalog during outage.” → RA-GRS storage or Cosmos DB.
-
Trick Q: “Availability Sets alone can protect against regional outage.” → False.