Last updated
The Load and Performance Testing feature in ratl.ai enables users to automatically generate and execute load tests to evaluate an API’s ability to handle high traffic. By leveraging AI, you can simulate real-world scenarios, measure performance under load, and optimize your API accordingly.
This feature supports various load generation formats, including cURL commands or postman collection, and can create tailored load tests to ensure API performance under stress.
Automated Load Test Generation: Automatically generate load tests based on cURL commands.
Customizable Load Conditions: Configure concurrent users, request rates, and test durations.
Execution Tracking: Monitor performance metrics such as response times, throughput, and error rates.
Detailed Results Analysis: Gain insights into how your API performs under different traffic loads.
Optimization Support: Identify bottlenecks and make improvements to enhance API performance.
Before using the Load and Performance Testing feature, make sure:
The API endpoint is defined and accessible via a specific cURL command or Postman exported collection.
You have access to ratl.ai and the appropriate permissions to generate and run load tests.
Navigate to the Load and Performance page.
Click on Add Simulations.
Provide the API specification using one of the following options:
Paste a cURL command.
Upload a Postman Collection (Postman collections must be v2.1.0, v2.0.0, or v1.0.0 and are limited to 10 APIs).
Select a Load Generator (currently supporting Gatling and JMeter, with future support for Locust and K6).
Review the pool injector's resource allocation:
Ratl Cloud Load Injectors:
Seamless load injectors hosted on Ratl’s in-house cloud infrastructure.
Offers a pay-as-you-go pricing model, calculated based on the load testing specifications or executions.
No need to manage any infrastructure—ideal for quick, scalable load testing setups.
On-Premise Kubernetes Load Injectors:
Users can integrate their own Kubernetes clusters to serve as load injectors.
Currently supports Kubernetes clusters, with Azure, GCP, and AWS support coming soon.
This option allows the Ratl injector pod instances to run directly within your infrastructure, reducing costs by leveraging your own resources.
For detailed steps on integrating Ratl with your Kubernetes cluster, refer to our guide: Integrate Ratl with Kubernetes Clusters.
Configure the injector configs with the following parameters:
Number of Users: Define how many concurrent users will be simulated during the test.
Max Duration: Set the maximum duration (in minutes) for the test.
Once configured, click Next to proceed for code review.
Tip: Ensure the API specifications are correctly formatted to avoid errors in test generation and Users cannot exceed the predefined limits of 10 instances or 4 CPU cores and 4 GB of memory.
For higher requirements, users can contact the admin to request an upgrade.
Navigate to the Code Editor after the load test scripts are generated.
Review the automatically generated load test scripts.
Use the prompt or the code editor to modify the scripts, if necessary.
Ensure that there are no issues in the script, as any errors will result in compilation failure and prevent the simulation from starting.
Review the script carefully, knowing that it is autosaved. You can return to the Simulations Listing page at any time to view or edit the code.
Once the code has been reviewed and confirmed, click Start Simulation to proceed.
Note: The complexity of the cURL command or Postman collection may impact the load test script generation time.
Once the simulation is started, you can either wait for the simulation to complete or navigate back to the Simulations Listing page.
If needed, you can stop the simulation at any time during execution.
Check the status of instances in the Simulation Flow screen, which will update based on the resources you configured (e.g., number of pods, CPU, memory).
Access the logs to monitor the simulation's progress in real-time.
Once all instances have completed, the Generate Report button will become enabled. You will not be able to generate a report until the simulation is fully completed.
After the simulation is complete, click on Generate Report to proceed to the analysis phase.
Note: Please be aware that you are limited to 20 simulation runs per day. If you reach this limit and require more executions, you will need to wait until the following day or contact an administrator about upgrading your quota. This daily cap applies to every user, regardless of subscription plan.
Review the results after the simulation execution to evaluate how the API performed under the load.
Analyze key metrics such as response times, throughput, error rates, and resource utilization to assess the API's behavior.
Use the detailed performance report to identify any bottlenecks or areas for optimization.
Share the generated report with stakeholders for further review and insights.
Pro Tip: You can share the results to generate comprehensive reports for stakeholders, making it easier to communicate findings and recommendations.
For any performance bottlenecks or failures, use the provided logs and performance metrics to investigate the cause.
Make necessary changes to the API or adjust the load test configuration to optimize performance.
Re-run the load tests to validate that the changes have improved performance.
ratl.ai will generate a new report, confirming the improvements or highlighting any persistent issues.
Your API’s performance under high load is thoroughly tested and evaluated.
All bottlenecks or performance issues are identified and addressed.
Detailed performance reports are generated and can be exported for future reference.
