Question

You manage a set of Azure SQL databases protected by an auto-failover group with automatic failover enabled. The application requires that regional failover happen automatically during a prolonged outage, but the business will not tolerate silent data loss during transient network blips. You configure the failover group's read/write failover policy to Automatic and set GracePeriodWithDataLossHours to 4. During a real regional outage where the primary region is unreachable and replication cannot catch up, what behavior will occur?

Accepted Answer

Failover is delayed for up to 4 hours to allow the primary to recover; if it remains unreachable after the grace period, automatic failover proceeds even though some recent transactions may be lost.

Answer

Automatic failover is triggered immediately with zero data loss because synchronous replication guarantees all committed transactions are copied first.

Answer

Failover never happens automatically and must always be initiated manually because a grace period was configured.

Answer

The grace period is ignored for read/write failover and only applies to the read-only listener endpoint.