Performance Counting

Objectives

• To find performance problems

• Consistent recording of time and resource data

• Transparency with respect to scaling and resource utilization

• Usability with low effort

• Analyzability by PLANTA, if possible without dispatch of user data (in the form of a database copy)

Configuration and Usage Hints

Configuration

• The counting of the performance data in

  DT324 Statistics

  can be activated as follows:

  • By clicking on the Performance Counting: DEACTIVATED button in the Performance History Analysis module.

  • By executing the performance_counting_enable() function via the Python API, or alternatively.

  • By setting the performance_counting parameter in the globals.conf configuration file and restarting the server subsequently.

• The advantage of the Python function is, that from the moment of execution, all started and running sessions are recorded.

• For time descriptive performance counting, the soft limit with respect to the record amount (@G80109) should be deactivated (for information on this, see Define Limits for Database Queries via Global Variables.

Tip

• With the following Python code, a macro module can be created which switches the performance history on when it is run for the first time and switches it back off in the second run.

  • After activating the performance counting, PCOUNT: ON will be output after activation of the performance counting, and PCOUNT: OFF.

'''Toggles performance counting globally and 
shows a message depending on current setting'''

cur_title = ppms.ui_get_system_title() 
on_str = 'Performance Counting: ON' 
off_str = 'Performance Counting: OFF'  

if ppms.performance_counting_is_enabled():     
    ppms.performance_counting_disable()     
    ppms.ui_message_box(off_str) 
else:     
    ppms.performance_counting_enable()     
    ppms.ui_message_box(on_str)

Details

• A status message upon start of a new session informs on the performance recording.

• Since the PLANTA Server writes a record for every event, measuring can reduce the system performance significantly.

  • It only makes sense to use it to detect errors or performance problems.

• If data is no longer needed, the DT324 Statistics table should be emptied in order to prevent the data storage of the database from being used excessively.

• Please make sure that performance history analysis is activated or deactivated globally for the entire system, not only for a particular session.

View the Data, Transfer to PLANTA

View/Interpretation

• The performance data can be displayed via the Performance History Analysis module.

• For an overview of the single data fields, see here.

• You have to consider that measurement uncertainties and rounding errors may cause negative values, like in the Unspecified (real) field.

Export for Transmission

• The Excel format should be used for data export:

  • For this purpose, open the above mentioned Performance History Analysis module.

  • Load all data by pressing CTRL + Q, or via the Data → Filter menu item.

  • Select the Extras → Send data to → Microsoft Excel menu item.

  • Select Bulk export.

  • Select the Statistics data area.

  • Activate the Activate "Split" Excel function for window 1 checkbox.

  • Start the export via Export.

  • After the progress dialog has been closed, an Excel window opens in which the exported files are located.

• The export result can now be saved and sent to PLANTA via e-mail.

• In addition to the performance counting, PLANTA requires information on the context of use of the PLANTA Server.

  • To do so, please fill and send the Performance Problems checklist.