The Performance Trend Graph (formerly called Time History graph) is a view of aggregated Web performance over time. It enables you to see trends as well as identifying time periods when performance was better or worse than normal. Time periods included are specified on the chart configuration page, and can be relative to the current time, or over a specific absolute time range that you select.
A special type of Performance Trend Graph, the Long Term Graph , is available for Transaction Perspective and Application Perspective measurements. This is an aggregated performance and availability graph covering up to two years of your measurements. See Long Term Trend Graphs below for details.
Performance trend graph
The Performance Trend graph page includes both the performance and availability graphs for one or more Keynote measurements, over a selected time period. When you create the Performance Trend graph from the Graph page, you can select whether to include a performance graph, an availability graph, or both, which is the default.
The Availability percentage is generated by dividing successful measurements by total measurements. This number can vary depending on whether you include Content Errors as Page Errors.
Performance Trend Graph: Top graph shows performance; bottom graph shows availability. Mouse over data points for data detail. Click data point to zoom in.
The data points on the performance portion of a Performance Trend graph represent an average of all agent response times measured at that time. For example, a data point taken at 10:00 a.m. represents an average of all the response times measured from all the reporting agents connecting to the site at that time.
The data points on the availability portion of a Performance Trend graph represent an average of the percentages of successful measurements (number of successful measurements divided by the total number of measurements). For example, an availability value of 97 means that 97% of the measurements for that time showed no errors.
The times listed on a Performance Trend graph are not actual times of day, but rather estimated times of day representing the measurement time intervals. For hour on a graph showing one-hour time intervals, 10:00 a.m. represents the measurement time interval from 10:00 a.m. to 11:00 a.m.
Errors excluded from availability calculation
Clicking any data point in the Long Term Trend graph drills down to a regular Performance Trend graph covering the time period included in that data point.
Unlike the other MyKeynote graph types, which cover a maximum of six weeks, Long Term Trend graphs can include up to two full years of data. Each data point in a Long Term Trend graph represents an aggregated “bucket” of data, which can be daily or weekly.
Long-term trend graph including one year of data
A long-term trend graph shows aggregated performance and availability data for an Application Perspective or Transaction Perspective measurement.
Long-term trend graphs
You can click on any column header to sort the table in either ascending or descending order based on the values in the column.
The new table tab displays the numerical values used to create the current performance trend graph.
You can select whether you want the Performance Trend graph to show performance, availability, or both (the default).
You can change the scale of the graph by specifying a different Y-Axis Min Value and Y-Axis Max Value.
You can change the graph selection criteria, for example by specifying a different time range you are interested in, or generate a different type of graph for this data set by using the controls beneath the Summary Performance Trend graph.
Change graph selection criteria
Before you can compare pages from multiple transactions, you must have the View only current Tests option selected in the > Settings > Charts page.
You can use a Performance Trend graph to compare performance and availability of different individual pages of different transactions. For example, you can compare the performance and availability of the first (home) pages of several of your transactions. To do this, select the transactions you’re interested in, and then select the specific transaction pages in the Transaction Page Selection menu before generating the graph.
Comparing pages from multiple transactions
Some internal errors are removed from consideration when computing Availability because they may be caused by problems within Keynote agents or infrastructure. To see a list of all Keynote internal errors that indicates which ones affect Availability, see Internal and External Error Code Listing.