Accelerating performance testing at Imagine Learning using AWS

Previously, Imagine Learning used a homegrown performance-testing process that required platform engineers to provision testing nodes, monitor the execution of custom scripts, and manage ongoing maintenance. Although its homegrown solution supported open-source tools, the platform team still needed to manage the infrastructure. As the demand for its applications grew, Imagine Learning needed to scale its resources to accelerate performance testing.

Imagine Learning started researching performance-testing alternatives for its next-generation internal developer platform. That’s when the company discovered Distributed Load Testing on AWS. The team quickly incorporated the production-ready solution into existing workflows without extensive setup or customization. “We didn’t need support to implement Distributed Load Testing on AWS,” says Blake Romano, Staff Software Engineer at Imagine Learning. “Out of the box, the tool has done a good job.”

Because the solution supports Apache JMeter, Grafana k6, and Locust, developers can run tests and collect results in a standardized manner by using familiar open-source frameworks. The solution can scale resources and simulate millions of transactions to meet the load demands of enterprise customers, automatically orchestrating distributed test execution across AWS Regions.

To validate performance, developers can launch a test on demand, schedule it, or trigger it programmatically through an API. The solution also automates test result collection, parsing, logging, and analysis, providing centralized visibility into latency, throughput, and error rates without additional tooling overhead. “A user can go into Distributed Load Testing on AWS and upload custom scripts that are built on Locust, k6, or JMeter to run simulations at scale,” says Romano.

Imagine Learning also used Distributed Load Testing on AWS to understand the resiliency of their applications by simulating traffic while intentionally taking some application components offline to confirm that recent fixes addressed past issues. These tests showed that the platform could sustain the expected performance even during disruptions, giving teams greater confidence in overall reliability. “Using Distributed Load Testing on AWS to check performance while disrupting parts of the system, we verified that a previous bug no longer occurred,” says Romano.

Imagine Learning estimates that it would’ve taken 2–4 weeks to build another in-house solution to support a single performance-testing framework. Instead, the company implemented Distributed Load Testing on AWS in hours—quickly improving scalability, accelerating insights, and enhancing the developer experience.

By automating infrastructure provisioning and test management, Imagine Learning reduced operational overhead and helped developers run performance tests more quickly and independently. A developer can now operationalize and run a custom performance-testing script in 2–3 hours.

As the learning solution provider incorporates new technologies into its products, teams can validate performance under variable conditions and confirm that new features meet scale requirements. “We can now focus on more important things, such as saving costs and improving application performance,” says Romano.