The `BatchManager` handles a structure that is basically a list of streams that will be used within a transaction.
If we add chunking, we have a more complex structure
This data (structure and streams) must be shared across Nuxeo nodes if we want the whole system to work across the cluster without having to enforce affinity.
The initial implementation being for JSF, this was not so much of an issue, but it is now critical since we are using this very same endpoint :
- for Nuxeo Drive
- for Rest API
This looks like a new use case for the TransientStore providing an abstraction API on top of :
- memory or Redis for the structure
- FS or S3 for the streams