Guidelines for Converting Siebel ARM File

Review the following information as recommendations when converting Siebel ARM files.

• Set the Siebel ARM granularity to level 1 for monitoring production deployments; set the granularity to level 2 for diagnostic purposes.

• Set the SARM Max Number of Files parameter to 0 in order to disable Siebel ARM file creation. This scenario can be useful when you are enabling Siebel ARM for use with other third-party response measurement tools.

• Make sure that the Siebel ARM feature has flushed data to the Siebel ARM file before converting the file. The Siebel ARM feature creates an empty Siebel ARM file before data is flushed to the file. For more information about this process, see the descriptions for SARM Data File Size and SARM Period in "About Siebel ARM Parameters and Variables".

• Change the value of SARM Memory Size Limit or SARM Period to a lower setting if the Siebel ARM files remain empty on a consistent basis. For more information about this process, see the descriptions for SARM Data File Size and SARM Period in "About Siebel ARM Parameters and Variables".

• Make sure that the Siebel ARM file name and path name, as necessary, are correct when you reference the Siebel ARM files in the commands.

• (Siebel ARM Analyzer Tool only) If the Siebel ARM Analyzer Tool cannot convert large Siebel ARM files or the output file is too large, then split the Siebel ARM file by using the -p flag with the tool. For more information about the -p flag, see Table 15-9.

• Concatenate Siebel ARM files to increase the amount of performance data for a given process. For example, as the Siebel ARM feature can save numerous Siebel ARM binary files for each process, concatenate these files to view performance data for multiple requests for this process. For more information about the number of files saved, see the description for SARM Max Number of Files in "About Siebel ARM Parameters and Variables".

  
  

  Tip: Use a third-party utility to concatenate Siebel ARM files on Windows. Use the command cat list_of_files > filename.sarm to concatenate Siebel ARM files on UNIX.

  
  
  
  

  Note: Only concatenate Siebel ARM files of the same process.

  
  

• Gather performance analysis data on your Siebel application before customizing the application. These baseline measurements provide a good reference when monitoring the performance of your Siebel application after any customizations.

• Run a user session trace analysis if there are performance problems for an individual user during a particular session. The user trace session data identifies each request that the user made and identifies which request required the longest time when compared to a base line.

• Use the performance aggregation data to diagnose performance at a given point in time or for a certain process. Reviewing the data by group can diagnose the area that is performing poorly. After reviewing a high-level view of the performance data, extrapolate a more detailed review by running the comma-separated value analysis. For more information about running this analysis, see "Running Siebel ARM Data CSV Conversion".

• Compile performance aggregation data over a period of time to determine a trend analysis.