Aurora Serverless PostgreSQL Architecture

Overview

This architecture documentation describes the CloudFormation stack that deploys an Aurora Serverless PostgreSQL cluster on AWS. Aurora Serverless is an on-demand, auto-scaling configuration for Amazon Aurora that automatically starts up, shuts down, and scales capacity up or down based on your application's needs.

Key Benefits

• Cost Optimization: Only pay for the database resources you consume, with automatic pause after inactivity
• Auto-Scaling: Seamlessly scales compute capacity from 2 to 4 ACU (Aurora Capacity Units)
• High Availability: Built on Aurora's proven architecture with automatic failover
• Data API: HTTP endpoint for running SQL statements without managing connections
• Fully Managed: No server provisioning or management required

Architecture Diagram

CloudFormation Stack

Infrastructure as Code Deployment

Environment Configuration

Environment:dev

Database Name:temporal

Master User:Mahira

Scaling Configuration

Min Capacity:2 ACU

Max Capacity:4 ACU

Auto Pause:300s

AWS Secrets Manager

Securely stores and retrieves database credentials (username and password) for the RDS cluster. Referenced dynamically during stack deployment.

Key Parameters:

Secret NameMahira

Fieldsusername, password

Aurora Serverless PostgreSQL Cluster

Fully managed, auto-scaling PostgreSQL-compatible database that automatically adjusts capacity based on application needs. Supports Data API for HTTP-based queries.

Key Parameters:

Engineaurora-postgresql

Engine Version13.9

Modeserverless

Data APIEnabled

EncryptionEnabled (at rest)

Backup Retention30 days

Auto-Scaling

Automatically scales capacity from 2 to 4 ACU based on workload demands

Auto-Pause

Pauses after 5 minutes of inactivity to reduce costs

Security

Encrypted at rest with automatic 30-day backup retention

Stack Outputs

Connection Information

• • Endpoint: DBCluster.Endpoint.Address
• • Port: 5432
• • Database: temporal

Resource Names

• • Cluster ID: dev-AuroraCluster
• • Stack Name: Exported for reference

Components

1. AWS CloudFormation

The infrastructure is defined as code using CloudFormation templates, enabling:

• Version-controlled infrastructure
• Repeatable deployments across environments
• Automated resource provisioning and updates
• Consistent configuration management

2. AWS Secrets Manager

Securely manages database credentials with:

• Encrypted storage of sensitive information
• Dynamic reference resolution during stack deployment
• Automatic rotation capabilities (optional)
• Access control via IAM policies

The template references secrets using the syntax:

!Sub '{{resolve:secretsmanager:$\{DBMaster\}:SecretString:username\}}'

3. Aurora Serverless PostgreSQL Cluster

The core database component with the following characteristics:

Engine Configuration

• Engine: aurora-postgresql
• Engine Mode: serverless
• Engine Version: PostgreSQL 13.9
• Compatible with: PostgreSQL wire protocol

Scaling Configuration

• Minimum Capacity: 2 ACU (Aurora Capacity Units)
• Maximum Capacity: 4 ACU
• Auto-Pause: Enabled (pauses after 300 seconds of inactivity)
• Resume: Automatically resumes when accessed

Security Features

• Storage Encryption: Enabled (at rest)
• In-Transit Encryption: Enabled by default
• Backup Retention: 30 days
• Network Isolation: Deployed in VPC (when NetworkStackName is configured)

Data API

• HTTP Endpoint: Enabled
• Use Cases: Serverless applications, Lambda functions, containerized apps
• Benefits: No persistent database connections required

Configuration

Parameters

The CloudFormation template accepts the following parameters:

Outputs

The stack exports the following outputs for use by other stacks:

• StackName: The CloudFormation stack name
• ClusterName: The Aurora cluster identifier (dev-AuroraCluster)
• DNSName: The cluster endpoint address for connections
• DBName: The database name (temporal)
• DBPort: The PostgreSQL port (5432)

Deployment

Prerequisites

1. AWS Account with appropriate permissions
2. Secrets Manager Secret created with database credentials:

   {
     "username": "your_master_username",
     "password": "your_secure_password"
   }
   

   CopyCopied!
3. VPC Configuration (if using NetworkStackName parameter)

Deployment Steps

Using AWS CLI

aws cloudformation create-stack \
  --stack-name aurora-serverless-dev \
  --template-body file://aurora-serverless-postgres.yaml \
  --parameters \
    ParameterKey=EnvironmentName,ParameterValue=dev \
    ParameterKey=DBMaster,ParameterValue=Mahira \
    ParameterKey=DatabaseName,ParameterValue=temporal \
  --capabilities CAPABILITY_IAM

Using AWS Console

1. Navigate to CloudFormation in AWS Console
2. Click Create Stack → With new resources
3. Upload the aurora-serverless-postgres.yaml template
4. Configure parameters as needed
5. Review and create the stack

Post-Deployment

After successful deployment:

1. Retrieve Connection Details:

   aws cloudformation describe-stacks \
     --stack-name aurora-serverless-dev \
     --query 'Stacks[0].Outputs'
   

   CopyCopied!

2. Test Connection (when cluster is active):

   psql -h <DNSName> -p 5432 -U <username> -d temporal
   

   CopyCopied!

3. Using Data API (from Lambda or application):

   const AWS = require('aws-sdk');
   const rdsDataService = new AWS.RDSDataService();
   
   const params = {
     resourceArn: 'arn:aws:rds:region:account:cluster:dev-auroracluster',
     secretArn: 'arn:aws:secretsmanager:region:account:secret:Mahira',
     database: 'temporal',
     sql: 'SELECT version();'
   };
   
   const result = await rdsDataService.executeStatement(params).promise();
   

   CopyCopied!

Cost Optimization Tips

1. Set appropriate auto-pause duration based on usage patterns
2. Use Data API to avoid connection management overhead
3. Monitor scaling metrics to optimize min/max capacity
4. Enable Aurora Serverless v2 for finer-grained scaling (requires template update)
5. Review backup retention period based on compliance requirements

Monitoring

Key metrics to monitor:

• ServerlessDatabaseCapacity: Current ACU allocation
• DatabaseConnections: Active connection count
• CPUUtilization: Processor usage percentage
• FreeableMemory: Available RAM
• NetworkThroughput: Data transfer rates

Access metrics via CloudWatch dashboard or AWS Console RDS section.

Related Resources

• Aurora Databases API — Create and list databases per client with region selection via API.
• Amazon Aurora User Guide
• Aurora Serverless v1 Documentation
• RDS Data API Documentation
• CloudFormation RDS Resource Reference