How to modify Runtime Adjustment of Workload Dimension when generating a model in AMG

Document ID:  TEC542007
Last Modified Date:  03/14/2016
{{active ? 'Hide' : 'Show'}} Technical Document Details


  • CA Capacity Manager
  • CA Capacity Management (CA Hyperformix)
  • CA Performance Optimizer


  • CA Performance Optimizer:Release:4.0.3
  • CA Performance Optimizer:Release:4.0.1
  • CA Performance Optimizer:Release:4.0




Client Workload dimensions are required to be constants because the client processes are initialized at model initialization (time zero) and changing them in the middle of a model execution would be confusing and ignored.


Note: This approach would also enable you to dynamically change the number of active client during a model execution but then you would have the issue of sorting out the down-stream (e.g., server) statistics with time-series or performance monitors.

If you generate the model in AMG, you can still manually apply this simple change to the generated AMG client behaviors.

Follow the below instructions:

  1. Attach each Client Workload type (different behaviors) to a workstation dimensioned some value greater than the expected maximum number of clients of that type.

  2. Disable "Use average statistics" on each of those workstations so that each one reports separate statistics.

  3. Define a global variable associated with each workload type (e.g., MaxCL1) which is initialized by your OnInitialize() function.

  4. Structure each client behavior similar to below (example attached):
    Integer index = computerGetIndex();
    if (index < MaxCL)
    <your client logic here>
    At the end of the model run you will need to ignore the client response time statistics with a value of zero.

Please help us improve!

Will this information enable you to resolve your issue?

Please tell us what we can do better.

{{feedbackText.length ? feedbackText.length : '0'}}/255


Not what you were looking for?

Search Again >

Product Information

Support by Product >


Join a Community >