Help

JBlitz Professional captures a wide variety of results related information as your test run proceeds. All this information is written to a test results file held on your hard disk. The file is XML format. The data saved with the test results comprises:

• Graphing data as displayed in any of the Graph windows.
• The log of the last 'n' errors as shown by the Error Inspector where 'n' is configurable.
• The log of the last 'n' requests as shown by the Response Recorder where 'n' is configurable.
• All statistics displayed by the Statistics Manager.
• Test case and summary feedback information shown in the 'Main Setup' tab of the main screen.
• Error indicators shown in the 'Main Setup' tab of the main screen and in the 'Test Results' screen.

Additionally, downloaded web pages, their headers and associated request messages can be saved to disk if a 'file-based' logging mode is chosen. The data is stored separately from the test results file itself, but the results file records the information about where to look to find the associated data.

You can switch between multiple sets of test results by using the functions 'Open Test Results' and 'Save Test Results As' under the 'File' menu. Only one set of test results can be loaded at a time.

There are a variety of ways that you can make your test results available to other applications. See the exporting test results section for more information.

JBlitz provides a set of tools that you can use to visually analyze and inspect your test results. These are available either whilst your test run is ongoing and data is still being collected (they're all updated every second or so), or once your test run has completed. 

The tools work off the current test results associated with your test configuration. You can compare results for different scenarios by loading up different results files into your configuration. Click on one of the links below for further information about the analysis tools provided by JBlitz.

 

Graphs	Graphical representations of how many hits and errors have occurred along with other information such as the download rate, connection and response times and responsiveness.
Error Inspector	Summary information in tabular format about errors that have been detected so far (if any).
Response Recorder	Summary information in tabular format about the latest requests made and responses received by JBlitz.
Statistics Manager	Summary statistical information about the test run such as the download rate, error rate, mean response time etc.
Request message, response page and response header information	The request and response messages for each record shown in either the Error Inspector or Response Recorder, or displayed on the right mouse popup menu against any web page icon.

Remember, especially for troubleshooting, that JBlitz also provides a debugger, extensive logging support, a Log Console for quick visual log feedback, and a couple of debug flags for issuing debug information to the standard output device.

Graph windows showing a variety of download information can be viewed whilst a test run is ongoing or after any run has ended. The graph works by gathering statistics for discrete time intervals as the download is occurring. The graph auto-scales to an appropriate scale and has various options for customization.

Currently the available graphs are:

• Hits. The hit count versus time
• Errors. The error count versus time.
• Transactions. Details the number of complete cycles that have occurred. A complete cycle is the downloading of all the web pages for a given test case.
• Download rate. The number of bytes downloaded over time. Includes the bytes contained in both the response page body and response header.
• Connection time. The mean time to connect for a given time period. This time is from the request for connection to the connection becoming established.
• Response time. The mean time to respond for a given time period. This time is from the time the request being issued to the whole response being received.
• Responsiveness. A measure of how well the web server is responding to requests. The responsiveness is calculated by taking the proportion of the entire total request-response time that is spent waiting to connect and download the response. The greater time spent waiting for either the connection or the response, the lower the responsiveness. It is an indicative measure only and can be affected by overloading of the client machine too. time

Graph windows can be configured to either automatically update the display every second or so, or to only update on a manual basis. This latter mode is appropriate if you want to take time to view graph information whilst a run is still ongoing. Note, however, that some operations on the graph will inevitably cause an update to occur - notably if a new graph line is added for which fresh information is required.

Click the 'Graph' icon on the main application toolbar or click the right mouse button on any of the test cases and select 'Graph...'. A window is shown displaying the download information graph for the current run or most recent run. Multiple windows can be displayed each with its own graph. Each window has a menu with an option to show a new graph. This new graph will have the same display options as the original graph.

The lines displayed by a graph window are configurable at the bottom of the window in the graph legend. You are free to add or remove lines. You can also change the test case or web page that a given line is displaying by clicking on the table cell in the legend and selecting another from the combo box. Similarly, the line color can be changed by clicking on the color cell and selecting another color from the combo box.

The line order in the legend determines the 'front to back' order on the graph itself. The topmost line is displayed at the front of the graph and the bottommost line is displayed at the back.

Currently graphs are available as mountain graphs, histograms (bar charts) or line graphs.

Tooltips are provided which show a summary of download information for each individual line in the graph (just hover the mouse over the line of interest).

You can optionally display the throttle value that was in force during the test run by ticking the throttle check-box. The throttle is shown as a solid black line and a scale is shown on the right hand edge of the graph from -100 (minimum throttle) to +100 (maximum throttle). If the throttle is being displayed, throttle values will be exported alongside the graph data values whenever you export the graph data.

The data points shown for any of the graphs can be exported using the menu option Export->Graph Data... which brings up the Export Graph Data dialog for the selected graph. Review the information displayed, then click the 'Export...' button to export the data set to a file of your choice using whatever item delimiter suits.

The Error Inspector window shows a list of errors for a particular test case, or all errors.

The Error Inspector allows you to quickly identify problems that have occurred during a run and see when and what happened. If the error is one of -

• A HTTP response code error.
• An error text found error. Either via the 'standard error text' or 'page-specific error text' mechanisms.
• A 'wait time' error.
• A custom error raised by a global response checker, a page-level response checker or a customization module.

- then you can see the returned page text and headers by using the right mouse button after selecting the error to inspect. The other error types, such as connection failed error, do not have any reply text to view because these errors indicate that no page could be downloaded.

Tooltips are provided to explain the error columns shown. Columns can be sorted by clicking on the column titles, and can be rearranged by clicking and dragging the column titles to the left or right. At the bottom of the screen is a test case chooser allowing you to refine the display to show only errors for a particular test case or, indeed, a particular virtual user within one of those test cases.

The returned error response and headers are viewable for each error using the right mouse button to invoke the popup menu. Note that you will need to tick the relevant preference settings in the Preferences dialog in order to instruct JBlitz to record this information. Note also that the error page and headers are not saved with the test results, so no error pages or headers will be viewable for test results that have been saved and then reloaded.

The Error Inspector automatically updates its display every few seconds to reflect any new errors that have occurred. Any errors displayed can be removed by using the right mouse menu. Once an error has been removed, it is lost permanently. Of course, the error will still be detailed in logging output made at the time.

Click the 'Error Inspector' icon on the main application toolbar or click the right mouse button on any of the test cases and select 'Error Inspector...'. The Error Inspector window is shown. A single Error Inspector window is maintained for the whole application.

The Response Recorder keeps track of the latest 500 requests and responses issued and received by JBlitz. This window details, for each request, the following:

• The request number. The request number is a system assigned number that can be used to determine the order in which requests were made. Note that this is not necessarily the same as the order in which the requests will have been received by the web server.
• The URL being queried including the full query string.
• The time the request was issued.

If a particular request has received a response, the following fields are also displayed:

• The time the response was received.
• The success or failure of the response. This is simply determined as whether the response has been deemed an error response (see the section on error detection for more).
• The HTTP response code.
• The length of the response page body.

Tooltips are provided to explain the columns shown. Columns can be sorted by clicking on the column titles, and can be rearranged by clicking and dragging the column titles to the left or right. At the bottom of the screen is a test case chooser allowing you to refine the display to show only requests for a particular test case or, indeed, a particular virtual user within one of those test cases.

The returned response page and headers are viewable for each request using the right mouse button to invoke the popup menu on a given request. Note that you will need to tick the relevant preference settings in the Preferences dialog in order to instruct JBlitz to record this information. Note also that the response page and headers are not saved with the test results, so no response pages or headers will be viewable for test results that have been saved and then reloaded.

The Response Recorder automatically updates its display every few seconds to reflect any new requests that have been made. Any requests displayed can be removed by using the right mouse menu. Once a request has been removed, it is lost permanently. Of course, the request will still be detailed in logging output made at the time.

Click the 'Response Recorder' icon on the main application toolbar or click the right mouse button on any of the test cases and select 'Response Recorder...'. The Response Recorder window is shown. A single Response Recorder window is maintained for the whole application.

The Statistics Manager window gives an overall summary of statistical information for the test results gathered so far. The information is presented in tables that are updated every second or so whilst a run is ongoing. At the top left of the screen summary information for the particular filter selection is shown including details of how many virtual users are configured for the test run and the run start and finish times.

Tooltips are provided to explain the table columns shown. Columns can be sorted by clicking on the column titles, and can be rearranged by clicking and dragging the column titles to the left or right. At the bottom of the screen is a test case chooser allowing you to refine the display to show only requests for a particular test case or, indeed, a particular virtual user within one of those test cases.

Click the 'Statistics Manager' icon on the main application toolbar or click the right mouse button on any of the test cases and select 'Statistics Manager...'. The Statistics Manager window is shown. A single Statistics Manager window is maintained for the whole application.

Request messages, response pages and response headers are shown in response windows. These windows show you, on individual tabs, the request message issued, the downloaded response headers and the downloaded response page.

The response page can be shown as either Plain Text, HTML or as HTML with Image Download where the originating URL is re-queried to display the resultant page. The text is searchable via the search box shown in the bottom right of the screen.

Options exist to save the information shown to disk, browse the originating URL and to copy the URL to the system clipboard so it can be used by other programs.

To bring up a response window:

• From the 'Main Setup' tab of the main screen:
  1. Click with the right mouse button on any given web page and select the 'Response Received' menu item. Select the request of interest from the sub-menu and click 'Show Response...'.
  2. Click the right mouse button on any 'X' error indicators showing beside a web page icon. Select the request of interest and select the 'Show Response...' menu item.

   

• From the Error Inspector and Response Recorder windows:
  1. Click the right mouse button on any given record and select the 'Show Response...' menu item.

The ability to review downloaded content and headers alongside the request messages issued relies in part on the message log settings in force when the test run took place. These settings are specified under the 'Logs' tab on the main screen. If a fast 'in-memory' logging mode is in force, then the response information shown by the response windows will be lost when the application exits. Choose the 'file based' logging mode to ensure that a persistent record of your requests and responses exists on disk and that the information they display will then be retrievable even after closing and re-opening JBlitz.

JBlitz provides export functions for most of the data that it displays:

• Graph line values can be exported from the graph windows. Select the menu option 'Export->Graph Data'. Throttle data will be included if the throttle line is showing.
• Combined statistical information can be exported from the Statistics Manager window. Use the right mouse popup menu on any of the tables and select 'Export Combined Statistical Data...'.
• Data displayed in any of the tables can be exported using the right mouse popup menu and selecting 'Export->Selected Rows...' or 'Export->All Rows...'. This gives you the ability to export:
  • response records shown by the Response Recorder.
  • error records shown by the Error Inspector.
  • statistical information shown by the Statistics Manager.
  • graph legend information for each graph line shown in the graph legend at the bottom of each graph window.
  • test schedule results data shown in the results table for any of the test schedules managed by the Scheduler.
  • all web page entries shown in the Web Page Catalog.

After selecting one of these options to export data, you are presented with the export dialog. This window allows you to choose the delimiter that will be used to separate each data item from each other, and whether or not to also export header information. Header information usually makes up the first line of the data export and is normally just descriptive text for the value in each column.

Click 'Export...' to proceed and then choose a file to export to. Your data will be saved to the file and can then be imported into another application. This is an ideal way to transfer information from JBlitz to, e.g. Microsoft Excel or similar.

It is also worth remembering that the JBlitz customization module API gives you an alternate, and possibly more comprehensive, access path for gathering results data from your test runs and integrating it with other applications in an ad-hoc fashion.