---
title: "Load Testing"
description: "Define reusable API load specifications, run them with explicit SLOs, and inspect results."
---

> Documentation Index
> Fetch the complete documentation index at: https://docs.heizen.work/llms.txt
> Use this file to discover all available pages before exploring further.

# Load Testing

Load Testing models traffic against an API. Create a reusable specification with
a base URL, variables, and endpoints, then run it with virtual-user and service
level settings.

![Heizen Studio Load Testing page showing reusable specs, endpoint counts, run history, and the Run Load Test dialog.](/images/heizen-studio/load-testing.png)

## Create a load-test spec

1. Open **QA Agent → Load Testing** and choose **New Spec**.
2. Enter a spec name and **Base URL**.
3. Add variables for values such as IDs or tokens; reference them as
   `{{name}}` in a path, header, or body.
4. Add endpoints with a name, method, path, optional weight, headers, and body.
5. Save the spec. You can import endpoints from an API specification when that
   option is available.

## Run the spec

Choose **Run** on a spec and set the values you need:

- virtual users;
- duration in seconds;
- ramp-up time;
- SLO p95 latency in milliseconds; and
- SLO error-rate percentage.

Leave a field blank to use the displayed default. Confirm the target and
variables before starting a load test.

## Inspect results

Open a run from history to review the overview, latency and throughput details,
errors, and whether the configured SLOs were met. Past results are retained when
a spec is deleted, so keep the run as evidence and create a new spec when the
test objective changes materially.

> **Load testing affects the target**
>
> Coordinate the test window with the system owner. Start with a low traffic
> profile and increase it deliberately; do not use production credentials or
> targets without explicit approval.

Source: https://docs.heizen.work/studio/qa/load-testing/index.mdx