Skip to main content

Run a load test in a pipeline

Last updated on

Add a Load Test step to a Harness pipeline to run an existing load test as part of your delivery process. The step references a load test you already created, and you pass its parameters as fixed values, runtime inputs, or expressions. This lets you validate performance on every deployment and reuse one load test across pipelines with different values.

Feature flag

Load Testing is currently behind a feature flag (CHAOS_LOAD_TESTING_ENABLED). Contact your Harness sales representative to enable it for your account.

Prerequisites​

  • Module access: Access to the Harness Resilience Testing module.
  • A load test: An existing load test to reference in the step. Go to Python, JavaScript, or Java to create one.
  • A pipeline: A pipeline with a stage that supports the Load Test step, such as a Deploy or Custom stage.
Measuring health under load instead

This page adds a load test to a pipeline you already have. To run a load test alongside a probe that watches the service while the traffic is applied, create a composite load test instead, and Harness builds the pipeline for you. Go to Composite load tests to compare the two.

Add a Load Test step​

  1. Open your pipeline and select the stage where you want to run the load test.
  2. On the Execution tab, select Add Step.
  3. In the Step Library, under Resilience Testing, select Load Test. The Configure Load Test panel opens.
  4. Enter the step details:
FieldDescription
NameA descriptive name for the step.
IdThe step identifier, generated from the name. Select the edit icon to change it.
Description(Optional) What the step validates.
Load Test ReferenceThe load test this step runs. Select Select Load Test to choose one.

You can add more than one Load Test step to a stage to run several tests in sequence.

Select the load test​

  1. In Load Test Reference, select Select Load Test.
  2. In the Select Load Test modal, search for a load test and select its card. Each card shows the engine (Python, JavaScript, or Java), the last run status, and when it was last updated. The Summary panel shows the test's users and duration.
  3. Select Add to Pipeline.
Create a load test inline

Select + New Load Test in the modal to create a load test without leaving the pipeline.

Configure tool inputs​

After you select a load test, the Tool Inputs section shows the parameters the test exposes. The inputs depend on the referenced load test and its engine. For example, a Python (Locust) test on Kubernetes exposes:

InputDescription
durationSecondsTotal run time of the load test, in seconds.
rampUpTimeSecTime to ramp virtual users from zero to the target, in seconds.
workerCountNumber of worker pods that generate load.
targetUrlThe base URL of the application under test.

Required inputs are marked with an asterisk (*).

For each input, select the pin icon to choose how to provide its value:

Value typeWhen to use
Fixed valueEnter the value now. It does not change at runtime.
Runtime inputLeave the value open and provide it when the pipeline runs. Harness prompts you for it before execution.
ExpressionReference a pipeline variable or a prior step's output, such as <+pipeline.variables.targetUrl>.

Use Runtime input to reuse the same load test across pipelines and supply different values per run. Use Expression to derive a value from earlier stages, such as the URL of a service you just deployed.

Apply and run​

  1. (Optional) Select the Advanced tab to configure a failure strategy and other settings.
  2. Select Apply Changes to save the step, then Save to save the pipeline.
  3. Run the pipeline. If any input is a runtime input, Harness prompts you for its value before the step runs. The step resolves the inputs and variables, then runs the load test.

Go to Analyze load test results to interpret throughput, error rate, response times, and threshold outcomes.

Next steps​

Last updated on