Achieving high availability for your financial data warehouse with Amazon Redshift multi-AZ clusters.

An Amazon Redshift cluster is provisioned in a single Availability Zone (AZ) of the particular region of your choosing. Since December 2020, the Amazon Redshift RA3 family offers the cluster relocation feature. Only the RA3 instance family supports separation of compute and storage resources that enables several key features including relocation since compute resources can now be relocated to a new AZ and leverage a consistent copy of the cluster’s data in Redshift Managed Storage (RMS). You can turn on and manage cluster relocation from the Amazon Redshift console, AWS CLI, and even Amazon Redshift API. The cluster relocation feature moves a cluster to another AZ in one step without requiring any application changes.

To turn on cluster relocation, define a subnet group that includes multiple Availability Zones. If Amazon Redshift identifies more than one accessible Availability Zone, Amazon Redshift automatically chooses from the list of accessible Availability Zones to relocate the cluster. After relocation is complete, you use the same endpoint to access the cluster. Amazon Redshift deletes the original cluster's compute resources and returns them to the resource pool. Also when the cluster relocation capability is enabled, Amazon Redshift service can relocate your cluster in situations where AZ level issues prevent optimal cluster operation. However, note that this cluster relocation might not be possible in all scenarios due to potential resource limitations in a given Availability Zone. If this happens, Amazon Redshift doesn't change the original cluster.

Enterprise customers, especially in the Financial Services Industry, running mission critical workloads require a data warehouse with fast failover times and simplified operations that minimizes impact to applications. So in November 2023 Amazon Redshift has announced the general availability of Multi-AZ deployments for RA3 provisioned clusters. Multi-AZ clusters support running your data warehouse in multiple AWS Availability Zones (AZ) simultaneously and continue operating in unforeseen failure scenarios. A Multi-AZ deployment raises the Redshift Service Level Agreement (SLA) to 99.99% and delivers a highly available data warehouse for the most demanding mission-critical workloads. A Multi-AZ data warehouse also maximizes query processing throughput by operating in multiple AZs and using compute resources from both AZ’s to process read and write queries.

Your Multi-AZ data warehouse is a collection compute resources deployed simultaneously in two Availability Zones. The compute resources deployed in the primary Availability Zone are referred to as primary compute and those in the secondary Availability Zones are referred as secondary compute. A Multi-AZ data warehouse can automatically recover without any user intervention during an unlikely event such as an Availability Zone or infrastructure failure. The recovery process involves failing over from primary compute to secondary compute and designating secondary compute resources as primary. Additionally, new secondary compute resources are provisioned in a third Availability Zone. The performance of an individual query performance will not change after a failover has occurred. The overall throughput of your data warehouse will be reduced for a short time due to unavailability of compute resources in one of the Availability Zones. However, Amazon Redshift will automatically acquire capacity in another Availability Zone to ensure the same data warehouse processing capacity is restored.

The difference between 99.99% availability and 99.9% availability means an additional 7 hours of acceptable downtime for that service everywhere. That breaks down further to 1 minute and 26 seconds of downtime everyday, or 10 minutes and 48 seconds every week. So leverage the multi-AZ feature of Amazon Redshift RA3 clusters to achieve high availability on your critical financial data warehouse.

From your existing clusters, choose the cluster you want to convert to a Multi-AZ deployment. From Actions, choose Activate Multi-AZ. The modification summary appears. Click Activate Multi-AZ.

From the AWS CLI, use the modify-cluster command and the multi-az parameter.

You can also create a new Multi-AZ cluster by restoring it from an existing snapshot.

You can also resize your Multi-AZ data warehouse and specify a number of nodes or node type that is different from the current configuration.