Forecast Engine as a standalone AI service

When executing the forecast engine in batch mode, for a large number of item-location combinations, we have been limited by the number of processors and threads the Java Virtual Machine (JVM) process is able to effectively manage.

To improve scalability and performance, the Forecast engine is changed to run as a micro-service which is called by SCP on demand. This support running in multiple threads across many dedicated hosts.

Note: By default, this feature is enabled. You are not required any new role or privilege access to use this feature.