Amazon Aurora Serverless: An Overview, & Use cases

 

Amazon Aurora Serverless: An Overview

Amazon Aurora Serverless is an on-demand, auto-scaling configuration for Amazon Aurora (compatible with MySQL and PostgreSQL). It automatically starts up, shuts down, and scales the database capacity based on your application's needs — with no manual provisioning of database instances.

 Aurora Serverless Versions

Version	Status	Notes
Aurora Serverless v1	Older, limited features	Good for infrequent/variable workloads
Aurora Serverless v2	Current & Recommended	Scales in milliseconds, supports Multi-AZ, Global DB, read replicas, and custom endpoints

Use Aurora Serverless v2 for production-ready applications.

 Key Features of Aurora Serverless v2

·        Granular Auto-Scaling: Instantly scales up/down in fine-grained increments (0.5 Aurora Capacity Units - ACUs).

·        Always On: Unlike v1, v2 doesn’t pause during inactivity.

·        High Availability (HA): Supports Multi-AZ deployments.

·        Global Databases: Aurora Serverless v2 supports global databases for low-latency cross-region reads and disaster recovery.

·        Custom Endpoints: Can use custom endpoints in Serverless v2 clusters.

·        Read Replicas: Supports Aurora read replicas (great for read-heavy apps).

 Use Cases

·        Variable or unpredictable workloads

·        Development, testing, QA environments

·        SaaS applications with multi-tenant architectures

·        Event-driven or infrequent workloads (e.g., report generation)

 Aurora Serverless vs Provisioned

Feature	Aurora Provisioned	Aurora Serverless v2
Instance management	Manual	Automatic
Scaling	Manual or auto (via replicas)	Automatic and fine-grained
Cost model	Per hour (per instance)	Per second (based on ACUs)
Cold start latency	None	Minimal in v2
Availability features	Full support	Full support in v2

  Pricing

·        Billed per second based on the Aurora Capacity Units (ACUs) consumed.

·        Separate charges for I/O operations, storage, backup, and data transfer.

 ACU = Combination of CPU + memory + networking managed by AWS.

 Creating Aurora Serverless v2 (Console or CLI)

AWS CLI Example:

# bash

aws rds create-db-cluster \

  --engine aurora-mysql \

  --engine-mode provisioned \

  --scaling-configuration MinCapacity=2,MaxCapacity=16 \

  --db-cluster-identifier twtech-aurora-serverless-v2 \

  --vpc-security-group-ids sg-12345678xxx \

  --db-subnet-group-name twtech-subnet-group

Note: In v2,  twtech uses engine-mode = provisioned with scaling config to enable Serverless v2 (confusing, but that’s how AWS implemented it).

 Limitations

·        Aurora Serverless v1 has high latency in scaling and lacks full feature parity.

·        Aurora Serverless v2:

o   Only available in specific regions.

o   Minimum 0.5 ACUs, no “pause” feature like v1.

o   Some features (e.g., global write forwarding) are still evolving.