In this step, you will take the Sequence created earlier and prepare it for orchestrated execution. There is a video at the bottom of this page, if you want to see it. You will also learn how to review and interpret the results that Performance generates after a load test has finished.
You can use the same sequence you created previously. There is no need to record a new one.
Create a new timeline and start editing it
To begin, open the section in your project where your sequences are stored. This is typically found inside a folder you created for your onboarding work. Once you are there, create a new timeline. Right-click in the tree and choose to add a timeline. A blank timeline will appear at the top of the screen. This is where you visually compose and control your load test.
The timeline works like an editing canvas. Each sequence you place on it becomes a track, and the timeline represents how virtual users will ramp up and execute the test over time.
Drag your sequence from the tree into the first track on the timeline. Performance will automatically configure it to run from zero to one hundred virtual users over two minutes. You can adjust the duration by dragging the edge of the sequence block. As you extend or shorten it, the virtual user minutes shown in the top right corner update to reflect the load your test will consume. For now, keep the duration at two minutes.
Design your desired load profile
You can shape the load profile by double-clicking on the horizontal line inside the sequence track. Each click adds a waypoint. These waypoints define how virtual users scale over time. For example, you might create a step pattern that reaches two hundred users toward the end of the run. If you need to refine the values, select any waypoint and enter the number you want directly. The timeline updates instantly as you make changes.
When your load profile looks right, choose the region from which the test should run. Use the region selector to decide whether the load comes from Western Europe, Southeast Asia, or any of the available locations. Regional selection matters when you want to account for network distance or simulate realistic user behaviour.
Choose regions per sequence and enjoy live results
With the region selected for all sequences and the load profile in place, the test is ready to run. Click Run and Performance will begin generating the configured traffic. As the test is running, the timeline becomes live. You will see the load curve, response time trends, and request rates evolve during execution.
Below the timeline you can inspect the incoming metrics. Performance shows the number of errors, fastest response time, average, median, 95th percentile, 99th percentile, and highest observed values. These metrics help you understand how your system responds under pressure. Higher percentiles are particularly useful for identifying slow outliers.
As the test progresses, the visual curves begin to paint a picture of performance over time. The broader curve represents response times in milliseconds. The thinner blue curve indicates requests per second. The movement of these curves reflects the load pattern you created on the timeline.
Review automatically applied Application Performance thresholds
On the left side of the graph you will notice a performance scale based on the AppDex framework. It highlights three user experience ranges: satisfactory, tolerable, and unbearable. When response times remain in the lower range, the experience is acceptable. When they drift into the upper bands, it signals that users are likely to notice degradation or encounter unacceptable delays. By default, these are set with values as:
-
Satisfactory when response is below 1 second
-
Tolerable when response is between 1 to 3 seconds
-
Unbearable when response is above 4 seconds
However, you can customize these as you prefer per sequence.
Once the test finishes, you can click on any point in the sequence track to highlight the corresponding data. Colour coding follows the AppDex logic, helping you quickly spot areas where performance declined. If any part of the application behaved poorly, it becomes visible immediately.
Review AI-generated analysis of the run
Scrolling further down reveals an AI-generated analysis of the run. This section summarises key findings, highlights patterns, and offers potential explanations and recommendations. It is designed to help you move from raw numbers to meaningful insight without manual interpretation.
Save it and move on to the next step
At this stage you have taken a recorded sequence, placed it into a timeline, configured a load profile, selected a region, executed the test, and reviewed both live and post-run results. You now have a complete initial view of how Performance orchestrates load and how it presents performance data in a clear and actionable way. It’s that easy.