Github user david-streamlio commented on the issue:
https://github.com/apache/nifi/pull/2493
I completely agree on your observation about having to land the entire,
unbroken file into NiFi's repo rather than streaming through memory is
inefficient. I will talk to the Pulsar team about getting a true streaming API
implemented. It is a great suggestion.
If we were to write such an API, do you have some examples of how we would
leverage it inside NiFi? I thought we had to generate a flow file for our
Consumer processor. Is there a way to hand off streams instead?
---
