Depending on the type of event that raises the alarm, you could create an Event rule (either an Event rate or condition) which checks whether certain conditions are met before raising another event which asserts the alarm you want.
This requires a customisation of the event which is described further on our docops site:
Working with Event Rules - CA Spectrum - 10.2 - CA Technologies Documentation
This is the easier approach how you can achieve the objective.
1. First of all, create a SpectroWATCH on the servers' Model Type that can do the CPU threshold monitoring as per your requirements and don't set it 'active by default'. When it is activated on a server model this SpectroWATCH should raise alarm when the high aggregate CPU utilization condition stays for more 4 hours. Refer to the following SpectroWATCH example (Fourth Usability and Testing Watch Scenario section) for more details.
Watch Type Examples - CA Spectrum - 10.2 - CA Technologies Documentation
2. Create CLI scripts to activate/deactivate the SpectroWATCH on target server models. Refer to https://www.ca.com/us/services-support/ca-support/ca-support-online/knowledge-base-articles.TEC1657405.html?intcmp=searc… .
3. Use cron or Task Scheduler to schedule the execution of the CLI scripts at the right timing, i.e. run CLI script to activate the SpectroWATCH on Friday and run CLI script to deactivate the SpectroWATCH on Sunday.
I hope this help.