Cisco Cisco Unified Contact Center Enterprise 9.0(1)

The Diagnostic Framework service executes at a Below Normal priority so as to avoid 
adversely impacting server/application performance while running.  

 

Some CPU-intensive APIs of the Diagnostic Framework first check the overall system 
CPU utilization value (%CPU), and do not start the request if the %CPU value is greater 
than a threshold value.  

These APIs are: 



  LogMgr commands 



  TraceMgr commands 



  ConfigMgr command 

 
There are a few registry keys that control this behavior. Look in the following Windows 
Registry Key: 

HKLM\SOFTWARE\Cisco Systems, Inc.\ICM\Serviceability\DiagnosticFramework 

CPUThresholdSample 

5 

To get a more accurate reading of the 
%CPU, multiple readings are taken. 

This value says how many samples 
should be read. 

CPUThresholdDelay 

2 

The number of milliseconds to wait 
between each sample taken. 

CPUThresholdPercent 

60 

The percent value to compare the current 
%CPU to. If the %CPU is greater than 
this value, the API cannot start, and 
returns an error telling the user that the 
server is too busy, and to try the 
command later. 

The Diagnostic Framework service is designed to handle up to 20 concurrent web 
requests. The system was tested under load to work with this configuration. However, 
due to special circumstances if you must lower the number of concurrent requests, then 
you can modify the value of 

maxConcurrentCalls

 property in the service configuration file. 

1.  Stop the Diagnostic Framework service. 

 

2.  Launch Notepad and open the file: 

<ICM_Drive>:\icm\serviceability\diagnostics\bin\DiagFwSvc.exe.config 

Tip: You may want to make a copy of this configuration file before making any 
changes to it.