Load Testing with Opkey

Load testing assesses an application's performance by applying load, which is either less than or equal to the desired load. With load testing, you can evaluate whether the latest infrastructure can support the application and its sustainability under extreme user load.

Various scenarios are simulated to test the system's behavior under different load conditions, including high concurrent users, numerous requests, and heavy network traffic. This helps identify an application's maximum operating capacity and potential performance bottlenecks.

Load Testing is crucial because of the following factors:

It replicates real-world user scenarios.
Load testing is essential whenever there are code modifications in the application that could potentially impact its performance.
Identifying areas where the system's performance may degrade under heavy load and making informed decisions about infrastructure upgrades.
Detect and resolve memory leaks that could lead to performance degradation, ensuring the program does not consume excessive resources during operation.

With Opkey you can check the load capacity of Oracle EBS application while performing operations. Opkey’s load-testing keywords integrated into test cases enable you to measure the response time of your EBS workflow transactions based on simulated user activity. Let’s proceed to see how you can achieve that.

As an example, we’re creating a Purchase Requisition on the Oracle EBS application, which is then approved, and a Purchase Order is generated using that requisition number.

Select + Test to create a new test case.

Fill in the details on the window that opens and click Create.

Click on the Record button and select the Oracle EBS recorder to record the test steps.

Select Browser from the options and click on Start Recording.

As soon as you start performing steps on the application, they begin to get recorded. Login to the Oracle EBS application.

Navigate to the Requisition tab under Purchase expenses.

Enter the details of the Number, Item, Category description, quantity, etc to create a purchase requisition.

Similarly, you’ll have to proceed to add the required steps to generate the requisition number.

Once all the steps have been recorded, save the steps.

The recorded steps are now saved in the test case. Click on the Open Performance Steps tab to view the performance steps.

Expand the steps to view the generated performance steps.

We’re performing Co-relation or Dynamic Chaining from the test steps.

For example, a requisition number is generated on a step and must be mapped while creating a Purchase Order (PO) on a different step.

As you can now view the performance steps, we’ll proceed to further enhance the steps by adding keywords.

Right-click to add a step at the beginning with the keyword OracleLT_ThinkTime and set the desired wait time. Enter a minimum and maximum value between 1 and 10. The system will then pick a value within this range and wait for that time.

Now, Save the test case.

Search for the keyword and add it.

On the added step, click on the Set Data button.

Edit the data of the step, add the required minimum and maximum values and save it.

Select and expand the next step. Move to the step with the Show Response button. This will show the generated requisition number.

On the window, view the requisition number.

Create an output by clicking the Output tab and entering the output information to be mapped.

Now open the test step on which you want to map the saved output.

Click on the Set Input Data button, corresponding to the selected step.

Click on the Data Drive under the Input tab.

On the next window that opens, click on the Advance Options tab and select the Add JSON Path arrow to enter the response in the JSON format.

The window of data handling opens where you need to click on the requisition number.

The JSON path gets generated when the value is clicked. Set the generated path, by clicking the Set button.

The value gets mapped, and the static part now becomes dynamic.

Now we’ll add Oracle EBS-specific keywords.

We have the keywords OracleLT_StartTransaction and OracleLT_EndTransaction. First, add OracleLT_StartTransaction, followed by OracleLT_EndTransaction. These keywords return the response time results taken to perform the given steps through API.

Click to add the keyword OracleLT_StartTransaction.

Search for the keyword and add.

Once added click to edit and enter the data as LOGIN.

Now add the next keyword as OracleLT_EndTransaction.

Now click to Save the test steps.

Click on the Run button to execute the workflow.

To run the test, you need the Oracle EBS Performance Test Plugin, which supports 30 runs at a time from a single agent. You can execute these runs by mapping the Global Data Variable with the test step on the test case.

Select from the options of Run Functional Test or Run Performance Test.

Select the Local Browser option.

To execute with multiple users, create a Test Suite and add the respective Test Case.

To execute with multiple users, register the agent in the admin panel. When you click the Run button, the test will be queued and begin with the selected agent when available.

The executions are performed in the backend through APIs.

Let's look at the Grafana dashboard to see how the load increases with more users. The dashboard shows the load on the Oracle EBS application during the purchase order generation workflow.

With the increasing users, the per-user agent load is leveling high.

Below we can view the Response Time Statistics of the actions that we are performing.

The maximum and the minimum time taken to perform a step is seen here.

This is how we understand the load status of the application on the Grafana dashboard.