[
https://issues.apache.org/jira/browse/NIFI-1423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15113771#comment-15113771
]
Joseph Witt commented on NIFI-1423:
-----------------------------------
Louis-Etienne thank you for contributing. I took a look at the thread and
didn't quite see the answer i was looking for. The question I have for this
particular idea is what is the case in which one would want to penalize this
relationship given that it is specifically for 'no retry' cases and thus you'd
likely not loop/hit that same endpoint again. Keeping in mind the idea for
penalization of flow files is to say 'there is something about this flowfile
and the thing we're interacting with that we reasonably believe will solve
itself over time'. If instead the problem is that data from that relationship
should simply be processed more slowly one good option to consider is to have a
ControlRate or other processor with a slower scheduling interval.
I'm not saying this patch is a bad idea just want to understand how you
envision this working and to make sure we dont' already have a good mechanism
for it.
Thanks!
> InvokeHTTP ability to penalize on No Retry
> ------------------------------------------
>
> Key: NIFI-1423
> URL: https://issues.apache.org/jira/browse/NIFI-1423
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Core Framework
> Affects Versions: 0.4.1
> Reporter: Louis-Etienne Dorval
> Priority: Minor
>
> InvokeHTTP currently penalize FlowFiles on {{Failure}} (Exception) and
> {{Retry}} (5xx) but not on {{No Retry}} (1xx, 3xx, 4xx)
> As there is probably a reason for not penalizing it, there could be a way to
> opt-in for that behavior rather than forcing it.
> My usecase is discussed a bit more in the mailing list:
> http://mail-archives.apache.org/mod_mbox/nifi-users/201512.mbox/%3CCADSFSqHzCjxNNzJdbBAPS2x_iALVWL6zwR%2B8_Ecg%3DvUqpw9EgQ%40mail.gmail.com%3E
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
