Adding Load Testing to Chaos Engineering

How can you make sure that your services function under a wide range of conditions?

For many teams, load testing is their go-to method for seeing how their applications respond when there is a surge in traffic or high variability. Tools like JMeter, K6, Gatling, and LoadRunner enable teams to simulate these traffic changes to test their systems.

While load testing is a useful tactic, applications in production often rely on complex systems and dependent systems. For example, how would your load tests be impacted if there were latency introduced?

Chaos engineering is all about testing systems proactively to see how they handle a wide variety of scenarios. By combining load testing into your chaos experiments, you added depth to your testing to truly see how your applications behave.

Why Combine Load Testing and Chaos Engineering

Production isn’t clean. Services restart mid-request. APIs slow down without warning. Systems scale unevenly.

If you run a load test, you are only look at a 2D view of your application. As load changes, how does your functionality change?

By zooming out and incorporating load tests into your chaos experiments, you can surge load while also making other changes. For example, you could add latency between services, stress CPU, or restart an instance. The possibilities are endless.

Steadybit lets you coordinate traffic surges and fault injection into one experiment, where you can observe application performance in real-time. With Steadybit’s strong observability integrations, you’ll also have clear visibility into the impact of your experiment.

Load Testing Integrations in Steadybit

You can trigger load from your existing tools, use your own test scripts, and collect results directly in the platform.

JMeter

Use the JMeter extension to run distributed load tests from within a Steadybit experiment.

• Upload your .jmx test plan into the experiment editor
• Execute it alongside your chaos events
• Review full logs and results after the run
• Run JMeter locally or trigger it externally via API for large-scale tests

K6

K6 tests can run either locally (within the extension) or via K6 Cloud.

• Upload your script or run the test through the API
• Combine it with failure conditions like latency injection or pod disruption
• Download all K6 logs and metrics post-execution for deep analysis

If you’re already using K6 in CI, this integration allows you to reuse your tests as part of resilience validation.

Gatling

The Gatling extension supports HTTP-heavy simulations under load.

• Use your existing Gatling scripts
• Trigger the test from within Steadybit or externally
• Capture full Gatling reports after the experiment completes

This setup helps catch subtle slowdowns or error rates introduced by infrastructure failure under load.

LoadRunner

You can coordinate OpenText LoadRunner scenarios with Steadybit experiments.

• Trigger Steadybit from LoadRunner using an API call
• Reuse existing LoadRunner test plans
• Inject failures during the test execution
• Analyze results using LoadRunner metrics plus additional logs from Steadybit

LoadRunner’s metrics will reflect how your system responded under simultaneous load and failure. Steadybit adds visibility into logs, monitoring systems, and other infrastructure signals.

What You Get with Load & Chaos Testing

Combining load testing and chaos lets you:

• Observe system behavior under real pressure
• Detect slowdowns, retries, and bottlenecks
• Validate whether fallback logic actually works
• Confirm that SLOs hold when failures stack

Final thoughts

If you already know how your system behaves during clean load, now it’s time to find out what happens when things go wrong.

Add your existing load tests. Inject failures. Watch what changes.

If nothing breaks, you’re doing it right. If it does, better to find out here than in production.