Aurora Replica Sizing vs. Source DB Sizing - Do they need to match?

Update Follow-up question from customer: Would it simply be a performance impact when failover happens (i.e. master fails and the highest failover replica gets promoted)? Or would it have any impact during normal replication?

Answer:

Since Amazon Aurora Replicas share the same data volume as the primary instance in the same AWS Region, there is virtually no replication lag. We typically observe lag times in the 10s of milliseconds. How far behind the primary will my replicas be?: https://aws.amazon.com/rds/aurora/faqs/

By design, the replication asynchronous. So, the replica lag varies depending on the rate of database change. That is, during periods where a large amount of write operations occur for the database, you might see an increase in replica lag. Because we have to send the same log changes (Redo Logs) to Read replica to update the cache. If there are too many changes you might see a lag than expected. This does not depend on the which instance class you are using for Read Replica.

Re-invent Video: https://youtu.be/duf5uUsW3TM?t=422

Now, by having the smaller instance class for read replica:

1.) You might see lot of queries might be doing physical read (Read IOPS increase) instead of logical read. 2.) You might see a performance impact when failover happens to a Read Replica because it has less resources (CPU, Memory, etc. )