How LoginRadius built a performance-first DevOps culture with Gatling

Challenges

As traffic grew, LoginRadius needed to ensure that critical identity and access management workflows like login, registration, and MFA remained fast and resilient—even under load. Load testing had to become systematic, reliable, and repeatable.

Before adopting Gatling, LoginRadius relied on a homegrown load testing framework built on JMeter. This solution served them well in the early days, but soon reached its limits:

• Scalability bottlenecks – Testing beyond 1,000 RPS became impractical
• Manual execution – Load tests weren’t integrated into CI/CD
• Maintenance overhead – Internal tools were difficult to scale and manage
• Testing gaps – Performance was only checked during major releases

We were using homegrown or open-source solutions that were hard to maintain. That pushed performance testing to the back burner; it only ran when absolutely necessary.

Kundan Singh, Director of Product and Engineering

Why Gatling

After evaluating multiple tools, LoginRadius chose Gatling to take their performance testing to the next level. Key reasons included:

• Ability to simulate 15,000+ RPS across real-world scenarios
• CI/CD integration with GitLab pipelines
• Support for Groovy and JavaScript SDKs to keep test scripts flexible
• Easier test maintenance and reuse across environments
• Future-proof architecture to support a shift from self-hosted to cloud

Any team aiming to deliver great service should make performance testing a mandatory part of their SDLC. It shouldn’t be an afterthought.

From open source to Gatling Enterprise Edition

LoginRadius started with Gatling open source, then moved to a self-hosted version of Gatling Enterprise Edition.

Eventually, they adopted Gatling Enterprise Edition to:

• Avoid infrastructure setup and upgrades
• Gain access to on-demand load generators globally
• Centralize results with shared dashboards and live reporting
• Empower QA to run and analyze tests without DevOps bottlenecks
• Benchmark horizontal scaling performance (containers/instances) across regions

The cloud version also made it easier to iterate quickly, reduce test setup time, and ensure everyone—from engineers to product managers—had visibility into performance metrics.

Results

Today, LoginRadius runs twice-monthly load tests covering various API flows, including edge cases like invalid logins and token expirations. Each test helps validate the platform’s ability to scale, perform, and stay cost-efficient under pressure.

• P95 latency reduced from 500ms to under 250ms
• Confidence in each release with tests baked into CI/CD
• Failures surfaced earlier, before code hits production
• Reduced performance regression issues in production by over 80%
• Performance testing became part of the sprint, not an afterthought
• The system can handle 10x expected production load with <250ms API latency.

Compared to the past options, Gatling Enterprise Edition has definitely helped us streamline performance testing. It gave us visibility into key metrics we can now measure and improve, which directly helps us enhance our product for our customers.

‍