[jira] [Updated] (EAGLE-526) aggregation framework for mr history jobs

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

     [ https://issues.apache.org/jira/browse/EAGLE-526?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

wujinhu updated EAGLE-526:
--------------------------
    Description: 
Please review https://github.com/apache/incubator-eagle/pull/419 first.

This is the second step of the aggregation framework. This step uses results of the job metrics by each minute that generated by the first step. 

The framework is a storm topology. The spout decides which time range of job metrics needs to be aggregated, divide the time range into several pieces(For example, each piece is one hour) and emits these pieces to aggregate bolts.
The aggregate bolts aggregate the time range piece of job metrics that the spout assigns to them.
In aggregate bolts, they fetch job metrics by eagle server api and write the aggregate results back to eagle server.

> aggregation framework for mr history jobs
> -----------------------------------------
>
>                 Key: EAGLE-526
>                 URL: https://issues.apache.org/jira/browse/EAGLE-526
>             Project: Eagle
>          Issue Type: New Feature
>            Reporter: wujinhu
>            Assignee: wujinhu
>
> Please review https://github.com/apache/incubator-eagle/pull/419 first.
> This is the second step of the aggregation framework. This step uses results of the job metrics by each minute that generated by the first step. 
> The framework is a storm topology. The spout decides which time range of job metrics needs to be aggregated, divide the time range into several pieces(For example, each piece is one hour) and emits these pieces to aggregate bolts.
> The aggregate bolts aggregate the time range piece of job metrics that the spout assigns to them.
> In aggregate bolts, they fetch job metrics by eagle server api and write the aggregate results back to eagle server.



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