[jira] [Comment Edited] (HADOOP-15231) WavefrontSink for Hadoop Metrics2

["Howard Yoo (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-15231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16365956#comment-16365956 ] 

Howard Yoo edited comment on HADOOP-15231 at 2/15/18 5:26 PM:
--------------------------------------------------------------

Right, extending and re-using the code is always a good option. However, if you take a look at the GraphiteSink, you'd see these issues:
 # Much of its internal variables are defined as 'private' members, making anything extending it hard to use those.
 # Log messages have 'graphite' which makes it harder for others to change that, unless overwriting pretty much everything.

So, I can extend the class as you suggested, but at the end, I would have to overwrite most of the existing methods anyway which will probably make it end up in a similar situation. So, generally agree that extending something is useful, but I don't see too much values in this case.

Also, I feel that I need to create a unit-tests for this. Will it make it much better if I also add unit test for Wavefront format? Makes sense, and seems it is something that's required also.


was (Author: howardyoo):
Agreed,

Will change the code to extend the GraphiteSink instead of copying it... Also, I believe I need to add unit tests as well - would it be better if I added the unit tests?

> WavefrontSink for Hadoop Metrics2
> ---------------------------------
>
>                 Key: HADOOP-15231
>                 URL: https://issues.apache.org/jira/browse/HADOOP-15231
>             Project: Hadoop Common
>          Issue Type: New Feature
>          Components: metrics
>            Reporter: Howard Yoo
>            Priority: Minor
>              Labels: metrics, wavefront
>
> Wavefront is a SaaS based large scale real time metrics monitoring and analytic system capable of monitoring many different source systems. There are several Sinks available in Hadoop in order to capture various metrics to external system. The Wavefront data format follows a similar format as that of Graphite, with the addition of native point tag support and source value. The details are outlined here: [https://docs.wavefront.com/wavefront_data_format.html]
> It would be greatly helpful for both Hadoop and Wavefront to have a native integration using  Wavefront Sink to send Hadoop related metrics to Wavefront Server.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org