You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "Sachin Pasalkar (JIRA)" <ji...@apache.org> on 2015/11/24 06:04:11 UTC

[jira] [Created] (STORM-1343) Add a matrices for Trident which actually counts number of messages processed.

Sachin Pasalkar created STORM-1343:
--------------------------------------

             Summary: Add a matrices for Trident which actually counts number of messages processed.
                 Key: STORM-1343
                 URL: https://issues.apache.org/jira/browse/STORM-1343
             Project: Apache Storm
          Issue Type: Improvement
          Components: storm-core
    Affects Versions: 0.10.0
            Reporter: Sachin Pasalkar
            Priority: Critical


If we keep trident topology running without pumping any messages in Kafka. Storm UI still show increased count. After some investigation we found one blog mentioning that its not actual processed message count  https://github.com/miguno/kafka-storm-starter/issues/5

As user, its very confusing what is this count and some time it gets misinterpreted. We have seen core storm is showing the count correctly. As trident is abstract level of core storm can't we use those matrices? 

We also found a blog where user can manually add the matrices to his code which will come up to the Storm UI. http://www.bigdata-cookbook.com/post/72320512609/storm-metrics-how-to 

Can we implement processed tuple matrices in trident directly? That will helful to end user in understanding what is topology doing & what needs to done to increase EPS of particular component. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)