Aurora Provisioned Scaling execution block
Category: Database scaling
When you switch Regions, your Aurora provisioned database in the destination Region may be running a smaller instance class than your source Region — leaving you with insufficient compute capacity to handle production traffic. The Aurora Provisioned Scaling execution block automatically scales the destination instance to match the source instance class, ensuring your database is ready to serve full production load the moment traffic arrives.
Key benefits
Automatic capacity matching: Region switch reads the source instance class and scales the destination instance to match, eliminating the risk of under-provisioned databases receiving production traffic after a failover.
Instance creation when needed: If the destination instance doesn't exist yet, Region switch creates it with the correct instance class.
Cross-family intelligence: When the source instance type isn't available in the destination Region, Region switch automatically selects an equivalent or larger instance type with the same or more vCPU and memory, so you don't need to maintain instance-type compatibility mappings yourself.
When to use
Any recovery plan where Aurora provisioned instances must be at production capacity before traffic shifts.
Active-passive Aurora Global Databases: Your secondary Region runs a smaller (cheaper) reader instance that must be scaled up before receiving write traffic.
Cost-optimized standby Regions: You intentionally run smaller instances in your standby Region to save costs, and need automated right-sizing during failover.
How Aurora Provisioned Scaling compares to alternatives
Without this execution block, customers must ensure destination database capacity manually or through custom automation before switching Regions.
| Approach | Pros/Cons | |
|---|---|---|
| 1 | Aurora Provisioned Scaling block | Fully automated, handles cross-family mapping, creates missing instances, integrated with Region switch orchestration |
| 2 | Manual scaling | Full control over timing and instance selection, but slow, error-prone under pressure, requires operator availability during incident |
| 3 | Scripted automation (Lambda/SSM) | Customizable logic; Must build, test, and maintain; not integrated with Region switch sequencing; cannot leverage native plan evaluation |
| 4 | Pre-provisioning (always match) | Zero failover delay. Doubles cost in standby Region; wasteful for active-passive architectures |
The Aurora Provisioned Scaling block is the right choice when you want automated, validated capacity scaling as an integrated step in your Region switch recovery plan.
How it works
When the Aurora Provisioned Scaling execution block runs during plan execution, Region switch scales the target instance to match the instance class of the source instance through the following sequence:
If the target instance exists but is not in an
availablestate, Region switch waits for it to become available before proceeding.If the target instance does not exist, Region switch creates it in the target cluster with the instance class from the source instance.
If the target instance exists, Region switch validates that it belongs to the expected cluster, then compares the instance classes.
If both instances are in the same family and the target is smaller, Region switch modifies the target instance to match the source class.
If the instances are in different families, or the target is already at a larger size, no scaling is performed.
If the source instance type doesn't exist in the target Region, Region switch selects another instance type with the same or more vCPU and memory (for both create and modify operations).
Region switch polls the target instance until it reaches
availablestatus, then marks the step as complete.
Note
Region switch only scales up. If the destination instance is already equal to or larger than the source, no modification is made.
Configuration
Important
Before you configure the execution block, make sure that the plan's execution role has the correct IAM policy in place. For more information, see Aurora provisioned scaling execution block sample policy.
To configure an Aurora Provisioned Scaling execution block, enter the following values:
Step name: Enter a name.
Step description (optional): Enter a description of the step.
Global cluster identifier: Enter the identifier for the Aurora global cluster.
Cluster ARN for
Region: Enter the Aurora database cluster ARN for each Region in the plan.Instance ARN for
Region: Enter the Aurora database instance ARN for each Region in the plan.Timeout: Enter a timeout value.
Then, choose Save step.
What is evaluated as part of plan evaluation
When Region switch evaluates your plan, Region switch performs several checks on your Aurora Provisioned Scaling execution block configuration and permissions. Region switch verifies that the following is correct:
Both instance ARNs are well-formed.
At least one instance exists.
Any existing instances belong to the expected cluster.
Both cluster ARNs are well-formed and exist.
Both clusters are members of the specified global cluster.
Region switch also validates that the plan's IAM role has the required permissions for Aurora provisioned scaling. For more information about the required permissions for Region switch execution blocks, see Aurora provisioned scaling execution block sample policy.
The correct IAM permissions are essential for the proper functioning of the Aurora Provisioned Scaling execution block. If any of these validations fail, Region switch returns warnings that there are issues, and provides specific error messages to help you resolve the permissions or configuration issues.
Related resources
Amazon Aurora DB instance classes in the Amazon Aurora User Guide
