Increasing Application Throughput by 10x Using Amazon Aurora Serverless with Alloy

Alloy provides banks and financial technology companies with configurable identity verification software and automated solutions for reaching credit decisions, managing fraud risk, and monitoring transactions. The company had been using Amazon Aurora PostgreSQL—a fully managed, PostgreSQL-compatible, and ACID-compliant relational database engine—for its database needs. As Alloy began talking with larger enterprise customers, the company wanted to scale to support much higher transaction volumes.

Alloy first worked alongside AWS solutions architects to set up a sandbox environment and test its capacity for handling enterprise-level transaction volumes, simulating a real-life customer environment. Using historical data, Alloy estimated the number of requests that larger customers might typically send and designed tests based on that load.

Alloy adopted Aurora Serverless because of its auto scaling capabilities, which would help manage the seasonal and time-of-day fluctuations in traffic volume that are typical in financial services. As part of its testing strategy, the company also implemented Amazon RDS Proxy, which makes applications more scalable, resilient, and secure, to manage database connections more efficiently and prevent connection overloads during high-volume periods.

“The AWS solutions architects were very helpful,” says David Blackman, staff engineer at Alloy. “We enacted almost all their recommendations and saw a huge jump in the total throughput we could handle, which was amazing.”

By adopting Aurora Serverless, Alloy transformed its database infrastructure to increase throughput while reducing latency. Alloy boosted the overall throughput by up to 10 times, processing high volumes of transactions per second with no performance degradation. The company can also automatically adjust capacity to maintain consistent performance during demand spikes, removing the need for manual capacity planning. In addition, Alloy adopted Amazon Aurora I/O-Optimized, a cluster configuration that offers improved price performance and predictable pricing for customers with I/O-intensive applications.

Thus, the company was able to manage its heavy read-write workloads more efficiently. Furthermore, Alloy has raised throughput for customers’ onboarding workflows by 5 times. "Aurora Serverless is a dominant solution in its class for maximizing throughput while minimizing the engineering hours we need to maintain it,” says Blackman. “That’s a perfect combination for where we’re at as a company—a fully managed service that also has high performance.”

A major advantage of Aurora Serverless is its ability to handle fluctuating workloads efficiently. Alloy can automatically scale database resources up during peak hours and down during quiet periods, optimizing costs. The company no longer has to worry about database restrictions or frequently revise scaling policies.

“By going serverless, we can be minimally involved with the resources we’re running, knowing that we can rely on fast scaling that’s completely seamless for us with virtually no manual intervention,” says Mariia Rudenko, engineering manager for infrastructure at Alloy.

“On AWS, we can spend less time on operations and focus the team’s energy on other initiatives,” says Rudenko. “The more initiatives we can do, the more the team can grow and improve other processes.” Alloy can now focus on business growth rather than infrastructure management, confidently onboarding new enterprise clients while maintaining consistent performance and reliability.

On AWS, Alloy has created a data layer that can ingest 3,200 events per second per customer, compute approximately 1,200 near real-time decisions per second, and onboard about 600 entities per second. As the company moves forward in its transformation journey, it will continue refining its database architecture to support ever-increasing transaction volumes. Alloy is also exploring the Aurora PostgreSQL scale-out capability that businesses can use to distribute a database workload over multiple Aurora writer instances for even greater scalability.

“We really appreciate the level of automation and quality that AWS provides and how the AWS team is very customer focused in everything it does,” says Rudenko. “Being able to work alongside AWS solutions architects and having all these services that support that same mindset has been critical for our business.”