# Amazon Aurora Serverless AAS (Average Active Sessions)

# Introduction

Average Active Sessions (AAS) in Amazon Aurora Serverless, particularly with v2, is a key metric for understanding database load and how the serverless configuration dynamically scales. AAS represents the average number of concurrent database sessions that are actively performing work (using CPU, I/O, or waiting for resources) over a given period. It is a measure of how busy the database is rather than simply the number of open connections.

# How Aurora Serverless v2 handles Average Active Sessions:

-   **Autoscaling based on Load:** 

    Aurora Serverless v2 automatically scales its capacity (measured in Aurora Capacity Units or ACUs) up or down based on various factors, including CPU utilization, memory utilization, network throughput, and crucially, database load, which is directly related to Average Active Sessions.

-   **Fine-grained Scaling:** 

    Unlike v1 which scaled in large increments, v2 scales in much smaller increments (as small as 0.5 ACUs) to closely match the workload's demands, including changes in AAS. This allows for more efficient resource utilization and cost optimization.

-   **Impact on Performance:** 

    A high AAS indicates a busy database. Aurora Serverless v2's ability to scale quickly in response to increasing AAS helps maintain performance and prevent bottlenecks. Conversely, as AAS decreases, the database can scale down, reducing costs.

-   **Monitoring with Performance Insights:** 

    You can monitor Average Active Sessions and other performance metrics using Amazon RDS Performance Insights, which provides a visual representation of the database load and helps identify performance bottlenecks.

-   **Optimized Reads and AAS:** 

    Features like Aurora Optimized Reads can significantly reduce database load and, consequently, the Average Active Sessions required to handle I/O-intensive workloads by leveraging local NVMe storage, leading to improved performance and potentially lower ACU usage.