[
https://issues.apache.org/jira/browse/MINIFI-112?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15977365#comment-15977365
]
Jeremy Dyer commented on MINIFI-112:
------------------------------------
[~d9liang] This JIRA has gotten a little messy. Which piece in particular are
you interested in working on? The actual UpdateAttribute part as the JIRA
describes or the expression language discussions we were having above? The
UpdateAttributes has an open PR if you interested in looking/changing that one
so you don't need to start from scratch. It is rather old but at least the
context might help.
https://github.com/apache/nifi-minifi-cpp/pull/37
> Support UpdateAttribute for nifi-minifi-cpp
> -------------------------------------------
>
> Key: MINIFI-112
> URL: https://issues.apache.org/jira/browse/MINIFI-112
> Project: Apache NiFi MiNiFi
> Issue Type: New Feature
> Components: C++, Core Framework
> Reporter: Randy Gelhausen
>
> nifi-minifi-cpp agents can generate multiple "streams" of flowfiles.
> For instance, to monitor a host, nifi-minifi-cpp runs nmon, ps, netstat, and
> gathers logfiles from applications.
> But, for a given flowfile, any downstream NiFi collectors wont have
> visibility into the originating hostname, nor metadata about which "stream"
> (ExecuteProcess(nmon), ExecuteProcess(ps), TailFile(app1), TailFile(app2))
> generated it.
> One solution is to use a separate InputPort for each stream. This works, but
> burdens both the team working on agent flows, and the team managing the
> collector- they have to be in concert.
> A simpler (better?) approach is to allow agent teams to tag flowfiles with
> differentiating metadata via use of UpdateAttribute.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
