Github user markap14 commented on the issue:
https://github.com/apache/nifi/pull/2475
@jtstorck I should probably have read through all the comments before
adding my own :) Sorry about that. I did notice though, that you have resources
for "DISK" and "NETWORK" but they are not used anywhere. I would imagine that
any processor that changes the content of the FlowFile would get a "DISK" one -
which is a very large number of them. And perhaps even processors that read the
content? I wonder if that's actually necessary. Since the Processor shows how
much data is being read/written in the 5 minute stats, I wonder if we could
just drop that? Similarly, I think that the NETWORK utilization may be kind of
inferred in most cases - any processor that interacts with an external service
is likely to have high network utilization. But not sure it makes sense to
label every single one of those. Would recommend that we either remove those or
add javadocs explaining when exactly we recommend using those annotations if we
are not going to use them for each processor that touches f
lowfile content / network.
---
