Documentation at http://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_Storage.html says

This storage type can deliver single-digit millisecond latencies, with a base performance of 3 IOPS/GB and the ability to burst to 3,000 IOPS for extended periods of time.

Assuming that I'm using a m3.2xlarge (largest currently possible), if I provision more than 1000G, will the instance get a baseline performance of more than 3000 IOPS or will the instance be capped at 3000? For example, if I provision 4000G, will the baseline be 12,000 IOPS? Note that my question is for general purpose SSD not PIOPS.


The following graphic describes what type of IOPS you can achieve using GP2 storage. The amount of IOPS is a function of the total amount of storage you have provisioned up to a cap of 10,000 IOPS as of the writing of this post.

With RDS there is an added complication of when you scale up the amount of storage RDS may add an additional drive that you cannot see so the underlying IOPS can be affected by this in a non-transparent way because your instance may be using multiple EBS drives and not just one.

enter image description here

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.