Case Study: Reducing AWS RDS Costs by 40% for a Fast-Growing EdTech Platform
- priyanka panwar
- Case Studies
About
A fast-growing EdTech platform delivering live and interactive learning sessions was facing rising AWS costs due to heavy reliance on Amazon RDS for PostgreSQL. With increasing user engagement and concurrent sessions across multiple regions, their monthly RDS spend had reached $8,000 USD, primarily driven by various high-capacity RDS instances (db.m7g.4xlarge and db.m7g.8xlarge) and their associated read replicas running full-time in Multi-AZ configurations.
Industries
- AWS, AWS Cloud Services, Elasticsearch, MongoDB, SquareOps
Share Via
Client Overview
A fast-growing EdTech platform delivering live and interactive learning sessions was facing rising AWS costs due to heavy reliance on Amazon RDS for PostgreSQL. With increasing user engagement and concurrent sessions across multiple regions, their monthly RDS spend had reached $8,000 USD, primarily driven by various high-capacity RDS instances (db.m7g.4xlarge and db.m7g.8xlarge) and their associated read replicas running full-time in Multi-AZ configurations.
The Challenge
- Multiple RDS PostgreSQL instances (db.m7g.4xlarge and db.m7g.8xlarge) with Multi-AZ enabled and dedicated read replicas
- RDS Multi-AZ setup improved availability, but standby instances were not utilized for reads, adding to the cost
- GP3 volumes provisioned with 6000 IOPS and 250MBps throughput, over-provisioned for actual usage
- No dynamic scaling based on traffic cycles (e.g., school hours vs late-night)
The Solution
We began with a targeted benchmarking phase and then moved into a structured production migration. The strategy focused on achieving similar or better performance using more cost-efficient resources:
🔍 Performance Benchmarking
- Compared:
- Existing RDS setup (db.m7g.4xlarge and db.m7g.8xlarge)
- Downsized but optimized RDS (db.r8g.2xlarge)
- Aurora PostgreSQL cluster (r6g.large writer + 1–2 r6g.large replicas)
- Benchmarked using pgbench and custom workloads with Performance Insights and CloudWatch metrics
- Results: While RDS r8g performed well and offered savings, Aurora PostgreSQL outperformed it in read scalability, failover speed, and storage efficiency
🚀 Migration to Aurora PostgreSQL
- Transitioned production workloads to:
- Aurora Writer: r6g.large
- Aurora Reader: r6g.large in a different AZ (for HA)
- Eliminated the need for a dedicated RDS standby by placing the Aurora replica in a separate AZ
- Leveraged the Aurora Reader Endpoint for reporting and analytics queries
- Configured auto-scaling for read replicas based on traffic load during peak hours
🔄 Additional Optimizations
- Scheduled automated stop/start for dev and staging environments using AWS Instance Scheduler
- Switched to default GP3 volume provisioning (3000 IOPS), saving ~$300/month
- Disabled Performance Insights post-tuning to reduce recurring monitoring charges
- Maintained rollback-ready snapshots during the cutover for safe fallback
| Setup Type | Instance Details | Estimated Monthly Cost |
|---|---|---|
|
Original RDS Setup
|
Multiple m7g.4xlarge + m7g.8xlarge, Multi-AZ + readers |
~$8,000 |
|
Optimized RDS Option
|
r8g.2xlarge with Multi-AZ + read replicas
|
~$5,700 |
|
Aurora Optimized Setup
|
Multiple r6g.large writer + r6g.large replica+ (HA)
|
~$4,800
|
The Results
✅ Reduced monthly DB spend by 40% while meeting performance SLAs
✅ Eliminated non-readable standby by leveraging Aurora’s cross-AZ replication
✅ Improved failover speed (~15s vs 1–2 mins in RDS)
✅ Achieved dynamic read scalability with fewer nodes
✅ Reduced over-provisioned storage cost by tuning GP3 IOPS
✅ Maintained service availability with zero application-level changes
Interested in similar outcomes?
We help teams evaluate and execute Aurora migrations with zero downtime and measurable ROI.
About Us
SquareOps is a one-stop solution for any enterprise, regardless of size or industry, that wants to grow by taking advantage of the benefits of cloud-native platforms. We are a team of passionate engineers helping businesses set up the right processes and culture of DevOps for different verticals. We help developers understand best practices for coding, security as a basic need, and building pipelines for easy deployments. Our service model feature is a big differentiator from our competitors in the market, as we offer services at affordable prices.
Related Posts
Case Study: Reducing AWS RDS Costs by 40% for a Fast-Growing EdTech Platform
- Case Studies
Smooth Migration of MongoDB & Elasticsearch to AWS
- Case Studies
Streamlining Deployments for Loconav with Automation
- Case Studies
Scaling DevOps & Performance for MobileSentrix
- Case Studies
Migration of MongoDB & Elasticsearch to AWS
- Case Studies
AWS Control Tower Strategy For EyeControl
- Case Studies