Failover keeps a service running by moving operation from an unavailable primary region or component to a ready secondary. In Azure, replicated data makes that handoff possible. section 10.5.1 section 11.4.2
Common Azure approaches include:
- Active Geo-Replication provides up to four readable secondary databases, but failover is manual. section 10.5.1
- Auto-failover groups automate the switch for Azure SQL databases and expose read-write and read-only listener endpoints. section 10.5.2
- Cosmos DB uses multi-region deployment, failoverPriority, and optional multi-region writes to continue serving global applications. section 11.4.2
A regional outage
Banning Books replicates its ordering database to another Azure region. An automatic failover group can promote the synchronized secondary when the primary region fails, reducing downtime without requiring someone to intervene immediately. Planned failover can synchronize data first, minimizing loss. section 10.5.2
C:/> az sql failover-group create --name banningbooksfailovergroup --resource-group BanningRG --server BanningBooksServer --partner-server BanningBooksSecondaryServer --failover-policy Automatic --add-db BanningBooksDB
The Azure CLI example shows how to create a failover group, name both database servers, select an automatic policy, and attach a database. section 10.5.2
