RDS Serverless v2, Postgresql 15.4 high idle cpu

1

I set up a database to evaluate RDS Serverless v2. However it looks like that the base cost is way too high to say it would be serverless service. Database is set up to 0.5 to 2 ACU's and it is constantly loaded at ~50% writer and reader CPU. While no traffic or connections. Simple projection shows that this would mean that by the end of month it costs ~250 € to just idle this database.

So the question is that is this the normal behaviour and cost for this "serverless" cluster?

asked a year ago251 views
1 Answer
1

RDS Serverless v2 operates with a minimum capacity threshold, which you've configured at 0.5 ACUs. This means you're continuously billed for this baseline capacity, regardless of database activity. Unlike some serverless computing services that can reduce costs to zero during idle periods, RDS Serverless v2 maintains a constant minimum resource level to ensure rapid response times and consistent database availability. The 50% CPU utilization you're observing could be attributed to various factors such as background processes, routine maintenance, or the specific way RDS calculates and reports utilization for Serverless instances. It's essential to recognize that Serverless v2 is optimized for specific use-cases, such as applications with variable workloads or multi-tenant systems. If your application has a steady, predictable workload, Serverless v2 may not be the most cost-effective solution, as it will consistently utilize CPU, memory, and IOPs resources.

When configuring your Minimum and Maximum ACUs, several factors need consideration. The AWS documentation provides detailed guidance on this: https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/aurora-serverless-v2.setting-capacity.html

It's crucial to align your ACU settings with your actual workload requirements. For instance, if your application demands a minimum of 2 ACUs but you've set the minimum to 0.5, the system will likely operate at 2 ACUs consistently to meet your workload needs. Therefore, carefully analyze your workload patterns and adjust your ACU settings accordingly to optimize performance and cost-efficiency.

profile pictureAWS
EXPERT
answered 6 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions