Github user phrocker commented on the issue:
https://github.com/apache/nifi-minifi-cpp/pull/295
@minifirocks per your comment twelve days ago you mentioned that you didn't
want to use a processor because you're putting this on every flow file. That's
pretty reasonable, but why do we need to use the attributes? This is something
that's intended for immutable information, right? could we not have a static
strings within the flow file for just the flow version that are serialized
along with attributes?
If we do that we can let the other information be sent via C2 and
correlated via provenance. I don't see the need to include meta information on
every flow file for information that's immutable. Mutable information is better
sent in the flow file because those immutable data points can be correlated
when evaluating flow files from a given source.
---
