Cloud Costs, Crushed: How a Mortgage Leader Slashed AWS SpendFintech
Our Client is a leading U.S.-based mortgage lender offering a variety of home financing solutions, including conventional, FHA, VA, and jumbo loans. They specialize in helping borrowers through the home buying and refinancing process with personalized service and a wide network of loan officers.
They approached us with a pressing challenge: their AWS development environment costs were steadily climbing despite stable business operations. Upon initial assessment, we identified that their infrastructure relied on statically allocated resources based on theoretical maximum capacity rather than actual usage patterns
challenge
- Static Infrastructure and Rising Cloud CostsTheir RDS, EC2, and EBS services were continuously running at full capacity regardless of demand fluctuations, resulting in:
– Significant resource underutilization
– Limited scalability capabilities
– Unnecessary expenditure during off-hours and low-demand periods - Autoscaling HurdlesThe Loan Data API module lacked autoscaling, causing:
– Performance dips during peak loads
– Wasted spend during idle times
solution
- Our team conducted a comprehensive analysis of the client’s AWS environment, examining billing reports, CloudWatch metrics, and service configurations to establish a clear performance baseline.
- This analysis revealed several inefficiencies:
– RDS CPU utilization averaging just 1.77%
– Average database connections at 2.2 – significantly below capacity
– Oversized RDS instances (db.m5.xlarge) consuming substantial resources while idle - Based on these findings, we developed and executed a multi-faceted optimization strategy:
- Right-sizing Recommendations– Implemented downsizing to T3 burstable instances where appropriate.
-Strategically deployed Reserved Instances for predictable, continuous workloads - Dynamic Resource Management– Engineered on-demand autoscaling for ECS in the development environment.
– Established custom CPU and memory-based thresholds calibrated through load testing - Intelligent Scheduling Framework– Built and deployed an automation system to safely shut down idle services outside business hours.
– Created logic for retention of critical services while scaling non-essential resources to zero - Governance and Best Practices– Implemented comprehensive tagging policies for resource tracking.
– Established backup processes and failsafes
– Deployed automated guardrails with alerting capabilities
– Developed clear documentation and communication protocols
Outcomes
- ~50% cost reduction in RDS Services alone
- 29% projected savings via Reserved Instance plan
- Optimal resource usage via auto-scaling of ECS for the Loan Data API
- Improved system performance and scalability
- Better DevOps collaboration and change management through:– Resource tagging
– Ownership mapping
– Transparent communication channels - Long-term cost savings through:– Usage pattern analysis
– Continuous optimization and scheduled audits
Let’s work together
Unleash your ideas, goals, and vision. Join us on the journey to
remarkable results.