[
https://issues.apache.org/jira/browse/NIFI-987?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14990299#comment-14990299
]
ASF GitHub Bot commented on NIFI-987:
-------------------------------------
Github user rickysaltzer commented on the pull request:
https://github.com/apache/nifi/pull/91#issuecomment-153848048
Thanks for all of the comments, I'll address them as soon as I get a
chance. As @apiri pointed out, If the connection fails, then we need to call
`client.close()` (if not null) in order for the Riemann library to clean up its
netty threads. Otherwise, it endlessly spawns threads which eventually causes
the OOM exception.
I don't particularly like the way I'm backing off when failing to connect,
as it seems to retry very quickly. This is why the OOM surfaces minutes after
killing Riemann. What is the best practice here for yielding for a longer
period of time?
catch (IOException e) {
if (client != null) {
client.close();
}
context.yield();
throw new ProcessException(String.format("Unable to connect to
Riemann [%s:%d] (%s)\n%s", host, port, transport, e.getMessage()));
}
> Add Processor for Writing Events to Riemann
> -------------------------------------------
>
> Key: NIFI-987
> URL: https://issues.apache.org/jira/browse/NIFI-987
> Project: Apache NiFi
> Issue Type: New Feature
> Reporter: Ricky Saltzer
> Assignee: Ricky Saltzer
> Fix For: 0.4.0
>
> Attachments: Sample Riemann Dataflow .png
>
>
> Riemann (http://riemann.io) is a new framework for monitoring distributed
> systems. It's particular useful for sending ad-hoc events such as, heartbeats
> and metrics. It would be nice if NiFi had a PutRiemann processor for writing
> events using the NiFi expression language.
> A simple use case would be a data flow that repeatedly checks specific
> services over TCP, HTTP, etc and checks into Riemann. Another example would
> be detecting a blip in events coming down a stream using a simple event
> heartbeat mechanism. I'll post a couple visuals to help make these examples
> more concrete.
> I have an initial PutRiemann processor made. I will post the patch via a
> Github pull request later today.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
