You could do two things. First one is to increase the size of vRAM assigned to vCSA to 17-18 GB, but if You have enough resources and it is going to demand a reboot of vCSA. Check if the problem is solved now.
The other choise is to increase a Java heap size for this vmware-vpx-workflow service. Follow this procedure:
VMware Knowledge Base
cloudvm-ram-size -l - shows memory allocation in MB for particular service.
cloudvm-ram-size -J service_name - displays the heap size
cloudvm-ram-size -C heap_size service_name - changes the heap size.
In this scenario reboot is not required.