Github user joewitt commented on the issue:
https://github.com/apache/nifi/pull/2475
I think my only concern is that as-is we're labeling a bunch of things as
"CPU" or "MEMORY" but not giving descriptions. As a user i'd see that and
thing 'well, how does this use memory'? For instance, does that mean each
flowfile's content is fully loaded in memory? Or does it mean part of one is?
Or all of a batch of them? Or if we say CPU usage for compression how should I
think about number of threads? Or in the case of compress content it might be
worth adding 'MEMORY" and explaining that it is actually really efficient and
can handle large objects without ever loading much in memory. So in that case
the resource consideration is to alleviate concerns. We're not qualifying the
usage consideration as good or bad in this approach. But merely "Hey here is a
resource usage consideration you should or might have in mind and here is how
this component works in that regard". Does this make sense? So, in that sense
I'd like to see us add descriptions to all these things w
e're tagging. Not saying it is a must for the PR but adding "MEMORY" without
explaining might just be alarming
---
