[
https://issues.apache.org/jira/browse/EAGLE-524?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
wujinhu updated EAGLE-524:
--------------------------
Description:
The users want to get the aggregation results of job counters group by some
dimensions(by cluster/cluster&queue/cluster&queue&user and so on).For example,
the Hadoop cluster administrators want to see the trend of HDFS_BYTES_READ in
the last three days group by each cluster.
To meet this requirement, I implement this by two steps. This PR is the first
step. The second step is contained in #430
In this PR, the history feeder generates counter metrics when it processed
history jobs. When the feeder gets one task of each job, it gets the counters
values and generate metric value in each minute by ( counter value / task
duration * task run time in each minute). So, we get counter metrics by each
job and in each minute.
> aggregation framework-job level metrics aggregation
> ---------------------------------------------------
>
> Key: EAGLE-524
> URL: https://issues.apache.org/jira/browse/EAGLE-524
> Project: Eagle
> Issue Type: New Feature
> Reporter: wujinhu
> Assignee: wujinhu
>
> The users want to get the aggregation results of job counters group by some
> dimensions(by cluster/cluster&queue/cluster&queue&user and so on).For
> example, the Hadoop cluster administrators want to see the trend of
> HDFS_BYTES_READ in the last three days group by each cluster.
> To meet this requirement, I implement this by two steps. This PR is the first
> step. The second step is contained in #430
> In this PR, the history feeder generates counter metrics when it processed
> history jobs. When the feeder gets one task of each job, it gets the counters
> values and generate metric value in each minute by ( counter value / task
> duration * task run time in each minute). So, we get counter metrics by each
> job and in each minute.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
