Creating a Test
Learn how to set up and configure a test in LoadForge.
LoadForge tests are designed to be repeatable and configurable for consistent performance testing of your service. You specify the host, load parameters, and target performance metrics, and can trigger a run at any time.
Test Examples
LoadForge provides several ways to create test scripts:
- Copy examples from our directory.
- Use the test wizard to generate a script automatically.
- Upload a browser recording in HAR format.
Creating a Test
To create a new test, go to the Tests menu and click “Create New Test”. You will be asked to provide the following key parameters:
Name of Test
A descriptive name to help you identify the test, such as “MyWebsite Submission API” or “Staging Server Load Test”.
Servers to Launch
The number of cloud instances (workers) LoadForge will create for your test. These are launched and destroyed automatically.
- Each worker can handle approximately 10,000 virtual users.
- LoadForge always launches a controller instance in addition to the workers.
Users to Simulate
The total number of virtual users across all test workers. These users simulate real traffic by making requests based on the defined test script.
- Users are divided among the worker nodes.
- User behavior is influenced by the
wait_time
setting in the test script.
Default wait time setting:
This setting makes each virtual user wait 5 to 9 seconds before making the next request, simulating realistic browsing behavior.
Spawn Rate of Users
How quickly users are introduced into the test. Scaling too fast may overload the server unrealistically. We recommend setting the spawn rate to 2-5% of the total user count per second.
Host
Select the validated host that the test will run against. The selected host determines the default base URL for all relative requests (e.g., /test
vs. https://your.host/test
).
If your test requires multiple hosts, ensure they are all verified. LoadForge will allow full domain names for any Hosts that you have added.
Region
Choose where to run the test from. This helps evaluate latency and performance from different geographic locations.
Available test regions:
- New York
- San Francisco
- London
- Amsterdam
- Singapore
- France
- Toronto
- Bangalore
Test Script (Locustfile)
Advanced users can customize their test script (Locustfile). The default script simply requests the index page of your site, but you can define user workflows, API requests, and custom scenarios. See our Locustfile documentation for more details.
Performance Targets & Test Passing Criteria
To help assess performance, LoadForge allows you to define pass/fail conditions based on:
Apdex Target Response Time (milliseconds)
Defines the response time threshold for an acceptable user experience. Lower values indicate stricter performance expectations.
P95 Target Response Time (milliseconds)
The response time under which 95% of all requests should complete. This helps detect outliers and performance degradation.
Error Rate Target (percentage)
The maximum allowed failure rate before the test is considered failed. Any errors beyond this threshold will trigger a failure notification.
These metrics ensure that test results are actionable by providing clear benchmarks for success or failure.
Running the Test
Once all parameters are configured, click “Run”. Before execution, LoadForge will:
- Show the estimated resource usage.
- Provide a cost estimate (if using your own cloud provider).
- Deploy the required servers automatically.
Next Steps
After the test completes, review the performance reports, including:
- Response times (P95, average, and worst-case).
- Throughput and request rates.
- Error distribution.
- Pass/fail results based on your configured thresholds.
By consistently running and refining your tests, you can improve performance, detect bottlenecks, and optimize your infrastructure effectively.