Analyzing Test Results: Intro to Performance Metrics
This video shows how to analyze test results. We will demo how to navigate to the summary, graph and detail view of the test result and how to open a sub-report of a page or transaction after the test is complete. You'll learn how to analyze a page with the maximum number of performance goal violations. You will also learn how to determine the number of users your website can handle with a few clicks on a graph.
website performance metrics, performance analytics,
average response time, load test result
This video concludes the Getting Started series.
After the e-Store test completed in the "run the test" video, a new tab with test results opened. Let's examine it. You can select one of several views from the toolbar. The first view is the "Summary" view. It highlights test run information and its main statistics. It also includes average performance metrics for: the entire test, test iterations, requests, transactions, and pages. For example, for pages, the Summary view displays: average response time, page request rate, number of pages and how many times they were requested, and how many errors, timeouts, and missed goals were encountered. It also displays a list of the slowest pages. Let's examine another view. The "Graph" view displays a large number of performance graphs. At the same time, it allows to quickly locate and highlight analytical information you're interested in. For example, I would like to analyze the behavior of a page with the most goal misses. The goal is the target response time which a page or transaction should not exceed. I’ll select the Pages graph from the dropdown and select one graph layout to maximize it. Then, I’ll go to the grid and sort by the “missed goal” column. The "HP Pavilion" page has the greatest number of the goal misses. Right-click and select "hide other curves" to focus just on this page. I would like to examine the area of the graph where the page is consistently slower than its goal. So I select it to zoom-in. Mouse-over any data point to display its value. As you can see, the graphs can be quickly customized to highlight areas of interest. If you prefer to analyze performance analytics in numerical form, select "Detail" view. Here you can display one or several grids with detailed performance information about: transactions, pages, virtual users, and other areas of interest. Besides providing the analytics, the test results help to answer specific questions about your website performance. For example, I would like to find out how many users the e-Store website can handle. On the Page grid, click on the average response time column to sort. The slowest page is "HP Pavilion" with an average response time of 2.8 seconds. Click the "Details" link to open the page sub-report in a new tab. You can open sub-reports of pages or transactions that you are interested to analyze in more detail. Like in the Test report, in the sub-report you can select one of several views from the toolbar. Click the "Performance" view. It displays page response time dynamics. Green, Red, and Purple lines display the page's average, maximum and minimum response times respectively. I will hide the maximum and minimum lines. The Blue line shows the number of virtual users. The horizontal line shows the response time goal, which is 4 seconds in this case. I click in the graph area where the response time consistently exceeds the goal. The red vertical line correlates response time to the number of users. Mouse over a point where the red line intercepts the users' line. The data point tooltip shows that when the number of virtual users reaches 42, the slowest page consistently misses its performance target. Now we know that the e-store website can handle 42 users. This was a brief introduction to StresStimulus performance analytics. In other videos we will review more topics about analyzing the test result such as pinpointing performance bottlenecks, analyzing errors and timeouts, and comparing multiple tests. Thank you for watching.