A high percentage of heap usage observed leads to slow/problematic working of Automation Orchestrator.
In the /services-logs/prelude/vco-app/file-logs/vco-server-app.log file, you see entries similar to:
Terminating due to java.lang.OutOfMemoryError: GC overhead limit exceeded
Environment
VMware Aria Automation 8.x
VMware Aria Automation Orchestrator 8.x
Cause
The vAPI plug-in introduces the opportunity of adding vAPI metamodels and endpoints. Each metamodel takes up sufficient amount of heap space. Too large amount of imported metamodels (more than 20-30) can cause problems with the appliance. The exact number depends on the overall heap usage by the Automation Orchestrator plug-ins and platform.
Resolution
To resolve this issue, remove unnecessary metamodels:
Download the attached vmware.support.package package.
Login to Orchestrator Client and go to Assets > Packages.
Click Import and select the downloaded vmware.support.package package.
Click Trust to trust the Publisher Certificate and click Import.
Go to Workflows and locate the imported workflow Remove resources.
Run this workflow and select the resources for removal.
Note: There are two options of executions depending on the use case:
Single Resource:
In the Choose resource box, type resource ID.
Select Remove resource? checkbox.
All resources in a category:
In the Choose resource category box, type Library/Vapi.
Select Remove metamodel type? to remove all VAPI metamodels.
Select Remove endpoint type? to remove all VAPI endpoints.
Note: It is enough to leave one metamodel per major vCenter version.
If there is a backwards compatibility in the API even leaving the vCenter with the newest version should be enough for the proper work of the plug-in.
Workaround:
If you cannot remove the metamodels, you can increase the heap memory limit.
Note: This is only applicable for standalone Automation Orchestrator instances and is not supported for embedded Automation Orchestrator instances in VMware Aria Automation.