Amazon Aurora Serverless v2 is an on-demand, auto scaling configuration for Amazon Aurora.
When using Aurora Serverless v2, you specify a capacity range (minimum and maximum ACU values) for each database (DB) cluster.
Each ACU combines approximately 2 GiB of memory, CPU, and networking resources.
The post is Part 1 of a two-part blog post series and focuses on understanding how certain database parameters impact Aurora Serverless v2 scaling behavior for PostgreSQL-compatible DB instances.
To help you understand, there are test cases run through test cases on the following parameters that are most often tuned.
When using default parameters and no additional features enabled, the ACU scales down to 0.5 (minimum configured).
Using a nondefault configuration for max_locks_per_transaction impacts ACUs not scaling down below 2, and this might vary depending on the value you pick for this parameter.
Based on our findings, we recommend using a minimum ACU of 2 or higher if your DB cluster is supporting a high-connection workload.
In Part 2 of this series, we review how the minimum and maximum range of ACU influences scaling behavior of Aurora Serverless v2, tailored to different business scenarios.
The post authors are Venu Koneru and Priyanka. They both work at AWS as Database Specialist Solutions Architects.