Multi-region configuration distributes a Cosmos DB account’s data across Azure regions. It can bring reads closer to global users and keep an application running when one region fails—but the extra copies come with extra bills. section 11.4.2 section 11.1.3
A practical configuration balances availability, performance, and operational cost:
- Add multiple regions so data can be replicated across locations and read latency can decrease for geographically dispersed users. section 11.4.2
- Use automatic failover to move service to another region when the primary region becomes unavailable. section 11.4.2
- Configure multiple read regions, and enable multi-region writes when applications need to write in several regions. section 11.4.2
- Set failoverPriority to control takeover order; priority 0 is highest. Add zone redundancy to protect against failures within a region. section 11.4.2
- Estimate storage, synchronization, and request-unit charges before enabling regions or autoscaling; replication can increase costs and write latency. section 11.1.3
An Australian primary with a U.S. secondary
An e-commerce account can use australiacentral as the primary region with failoverPriority=0 and eastus as the secondary with failoverPriority=1. Setting isZoneRedundant=True for both regions adds zone-level protection. section 11.4.2
C:/> az cosmosdb update --name alpacacosmosdbaccount --resource-group AlpacaRG --locations regionName=australiacentral failoverPriority=0 isZoneRedundant=False regionName=eastus failoverPriority=1 isZoneRedundant=False
Use the Azure CLI listing to inspect the region order and zone-redundancy flags for a Cosmos DB account. section 11.4.2
