How to decide on Xmx for heap size for a bulk ingestion project

Question

I am working on a bulk ingestion project for about 5 million documents (sized between 200KB-3MB each) to Alfresco using REST APIs and Spring Boot. My fat jar will runs like