Amazon OpenSearch Service now offers managed cross-Region disaster recovery that continuously replicates a primary domain to a secondary domain in another Region. You pair domains with one API call or a few console steps, achieving an RPO under one minute and an RTO under 20 minutes. Planned switchovers swap roles with zero data loss; unplanned failovers promote the secondary with loss limited to replication lag. The feature requires OpenSearch 3.7 or later, works within a single AWS account, supports customer-managed KMS keys, and carries no extra charge beyond standard secondary domain and cross-Region data transfer costs.


How managed disaster recovery works
After you create OpenSearch domains in two commercial AWS Regions, you pair them using the AWS Management Console or a single API call. The secondary domain stays continuously in sync with the primary, replicating indexes, mappings, and settings without any custom pipelines or scripts. Amazon CloudWatch metrics expose replication health and lag so you can verify the link before you ever need it.
For planned events such as maintenance windows or Region rotation, a switchover exchanges the primary and secondary roles with zero data loss. If the primary becomes unavailable, a failover promotes the secondary; any data loss is bounded by the replication lag at that moment. After a failover, you restore protection by pairing the new primary with a fresh secondary domain.
Managed disaster recovery requires OpenSearch 3.7 or later and supports domains encrypted with customer-managed AWS KMS keys. Domains using UltraWarm storage or OpenSearch Optimized instances (OR1 or OR2) are not supported. If you are on an earlier version, clone-based upgrades let you test OpenSearch 3.7 against your workload, cut over when ready, and roll back if needed.
Cost, availability, and operational notes
There is no additional charge for the managed disaster recovery capability itself. You pay standard rates for the secondary domain and for cross-Region data transfer between the paired domains. Pricing details are on the Amazon OpenSearch Service pricing page, and the feature is available today in all commercial AWS Regions where OpenSearch Service runs.
Because replication runs continuously, the secondary domain consumes compute and storage resources equivalent to the primary. Plan capacity for both domains accordingly, and monitor CloudWatch metrics such as replication lag and sync status to catch issues early. The single-account requirement means cross-account DR setups still need a custom approach.
Clone-based upgrades provide a low-risk path to OpenSearch 3.7: you clone the production domain, upgrade the clone, validate your workload, then promote the clone. This avoids in-place upgrade risk and gives you a rollback point if the new version behaves differently.
What this means for search and analytics teams
Teams running mission-critical search, AI retrieval, or log analytics workloads can now meet strict availability and business continuity requirements without building and maintaining their own cross-Region replication. The managed model removes the operational burden of custom scripts, snapshot scheduling, and manual failover runbooks.
Sub-minute RPO and sub-20-minute RTO put OpenSearch Service in line with database-tier DR expectations, which is significant for workloads that feed real-time dashboards, fraud detection, or generative AI pipelines. The ability to switchover for planned maintenance also eliminates downtime windows for version upgrades or Region rotations.
The constraints — OpenSearch 3.7 minimum, no UltraWarm or Optimized instances, single account — are important to factor into architecture reviews. If your domains rely on those unsupported features, you will need to evaluate migration paths or retain custom DR until support expands.
What to do next
Next steps: verify your domains run OpenSearch 3.7 or later, confirm they do not use UltraWarm or OR1/OR2 instances, then pair a primary and secondary domain in the console or via API. Enable CloudWatch alarms on replication lag and rehearse a switchover during a maintenance window to validate the runbook.
Source: Amazon OpenSearch Service announces managed cross-Region disaster recovery (AWS).



