setup: aws rds mysql aurora
Our DB has a writer and a reader instances. If we use the reader instance(which has a separate CPU) will large read requests have any affect on the Database that the Writer instance is connecting to?
Basically, just thinking worse case scenario. Want to let a new DA do new analytics work but want to be careful that there is no affect on the DB. Believe that both instances connect to the same DB but have separate CPU’s. So this should be fine. (as long as the reader is not used in other ways)
Since the two instances have separate CPU’s and the reader instance is not used in other ways. This should be fine with no affect on the DB.
2
Answers
No. Your usage of the Read Replica will not impact the performance of the Writer.
The data is replicated between the nodes at the storage level without impacting the Reader or Writer nodes.
No, there will be no impact on the writer. The data in the cluster storage volume is represented as a single, logical volume to the primary instance and to Aurora Replicas in the DB cluster. Aurora automatically maintains 6 copies of your data across 3 Availability Zones, with a write set of 4 and a read set of 3.
Hence it is ideal to use together with Reader Autoscaling and you can even combine it with RDS-Proxy to have a single reader endpoint for your application. Reader endpoints for RDS Proxy
If you don’t use serverless, you can lookup the io limitations per reader instance type here, but don’t confuse this with the expected io bandwidth. This figure refers to I/O bandwidth for local storage only within the DB instance. It doesn’t apply to communication with the Aurora cluster volume. https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Concepts.DBInstanceClass.html#Concepts.DBInstanceClass.Summary
Be aware of the IO-Optimized option, depending on your requirements. https://aws.amazon.com/about-aws/whats-new/2023/05/amazon-aurora-i-o-optimized/