Which solution will accomplish the move with the LEAST downtime and the LEAST application impact?
Modify the existing DB cluster and update the Aurora configuration to ג€Serverless.ג€
Create a snapshot of the existing DB cluster and restore it to a new Aurora Serverless DB cluster.
Create an Aurora Serverless replica from the existing DB cluster and promote it to primary when the replica lag is minimal.
Replicate the data between the existing DB cluster and a new Aurora Serverless DB cluster by using AWS Database Migration Service (AWS DMS) with change data capture (CDC) enabled.
Explanations:
Aurora Serverless cannot be enabled on an existing provisioned Aurora cluster. Modifying the existing cluster configuration to switch to Serverless is not possible without creating a new cluster.
While creating a snapshot and restoring it to a new Aurora Serverless DB cluster can transfer the data, this method would involve downtime during the snapshot creation and restore process, which is not ideal for minimizing downtime and application impact.
Creating an Aurora Serverless replica from the existing DB cluster is not possible since Aurora Serverless cannot act as a read replica of a provisioned Aurora cluster. This option does not provide a feasible migration path.
Using AWS DMS with change data capture (CDC) allows for continuous data replication with minimal downtime. It enables real-time data transfer while the application continues to function, ensuring that the new Aurora Serverless DB is kept up to date until the final cutover.
I have a feeling that the answer is:
Replicate the data between the existing DB cluster and a new Aurora Serverless DB cluster by using AWS Database Migration Service (AWS DMS) with change data capture (CDC) enabled.