Using the Performance Analysis Tool

The Performance Analysis Tool can be used if there is a performance issue identified for a specific story. It can be used to analyze single user workflows and break down their total end-to-end runtime. It is possible to compare story performance between users and over time.

The tool is a predefined story that identifies processing time for a story by frontend, network, and backend times for a specific session.

The tool is integrated with the System Overview, allowing for a more transparent value table for selecting the user action to be analyzed. and the Story Loading Sequence in the detail area of the Performance Analysis Tool.

You can launch the tool by going to System → Performance → Analysis Tool as shown here.

After you have launched the Performance Analysis Tool, you can filter by session day, session date, user, or by story (resource) name.

1. Set your filters to define the search criteria.
2. Select the item to analyze.

The entry page gives you a fast overview of the workflow that you plan to analyze, how many widgets have been rendered by the actions in the workflow, and what layer the most time has been spent in.

In the Details panel, there are three tabs on the entry page, allowing you to further investigate:

• Page Load Time
• Widget Drilldown
• Loading Sequence

The Page Load Time shows the real end-to end time, from triggering the action until the page has been fully rendered. This total time can be split into Page Preparation and Widget Load Time.

• Page Preparation: The time from the action start time stamp until the first widget starts loading.
• Widget Load Time: The time from the start of the first widget until the last widget has finished rendering.

In the following example, look at the values in the Total End to End Time and Maxima per Processing Layer chart and notice the widget load. Now, look at the Top 5 Widget - Total Widget Time chart. You can see that the main cause of this time is the P00M_OpIncFcst widget.

Scroll down to view Total Time per Date and Runtime Split per Date charts, where you can compare the behavior to other users and time.

The Widget Drilldown drills into information on single widget level per action, model, or page. Clicking on a widget in the Per Widget End to End Time table opens Widget Details. Here, you can view the details of each widget individually with information about the Backend Time, Network Time, and Frontend Time.

If you select a widget, for example, P00M_OpIncFcst , then you can view the details for that particular widget. To return, simply click the Clear Filters button.

In the Widget Drilldown tab, you can scroll down to view Widget Time per Date and Runtime Split per Date charts, where you can compare the behavior to other users and time.

Loading Sequence shows the aggregated widget and processing layer times by the start time, with information of corresponding action and story page.

In the Loading Sequence tab, you can also scroll down to view Widget Time per Date and Runtime Split per Date charts, which you can see at the bottom of the example above. Here, you can compare the behavior to other users and time.

Once you have analyzed the different elements of the story performance and you are ready to optimize your story or investigate further, you can open the story directly from the tool.