Application Connections after PITR Aurora Postgres or DynamoDB

Question

Hi! First question on this forum and it may seem trivial, but I'm trying to get around seamless reconnections of applications and interfaces after Point in Time Recovery. PITR of Aurora Instances and DynamoDB tables always result in new instances of either being created. Potentially these could be hosting connections from many targets. My experience (more that 2 decades) is with the likes Oracle and SQL Server and in situ restores/recoveries so no need to repoint applications. How are people handling the change of database target if they have to performs such a recovery?

(We are thinking of going RDS rather than Aurora because we can understand the process there, but this seems like a poor reason to choose one over the other).

Thanks, John

(We are thinking of going RDS rather than Aurora because we can understand the process there, but this seems like a poor reason to choose one over the other).

Thanks,
John

Answer

I can answer in terms of DynamoDB but not the relational tables as they both use a different means of connections. For DynamoDB its common to use [Parameter Store](https://docs.aws.amazon.com/systems-manager/latest/userguide/systems-manager-parameter-store.html) to store the table name. Your application would always retrieve the table name from the parameter store, meaning if you need to restore the table to one with a new name, that just means you need to update one value in parameter store and your applications will automatically resolve the correct table name.

Application Connections after PITR Aurora Postgres or DynamoDB

Relevant content