[
https://issues.apache.org/jira/browse/FLINK-1502?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15150166#comment-15150166
]
Maximilian Michels commented on FLINK-1502:
-------------------------------------------
Hi [~eastcirclek]! In addition to what [~jgrier] wrote:
The cluster setup should be the normal use case for Ganglia/JMX monitoring.
When multiple task managers run in the same JVM or on the same machine, the
reporting should work similar. Just namespace the task manager metrics. Task
managers already have an {{InstanceID}} which uniquely identifies it. On
clusters you normal wouldn't run multiple instances on a machine and its ok if
the output on Ganglia is not optimal then.
The metrics shouldn't go through the JobManager to be reported. Actually, they
already go there but just for displaying them in the web interface. However,
cluster tools should monitor processes directly at the nodes.
Metrics shouldn't be aggregated or combined. Users should be able to monitor
nodes and also identify differences in resource consumption.
> Expose metrics to graphite, ganglia and JMX.
> --------------------------------------------
>
> Key: FLINK-1502
> URL: https://issues.apache.org/jira/browse/FLINK-1502
> Project: Flink
> Issue Type: Sub-task
> Components: JobManager, TaskManager
> Affects Versions: 0.9
> Reporter: Robert Metzger
> Assignee: Dongwon Kim
> Priority: Minor
> Fix For: pre-apache
>
>
> The metrics library allows to expose collected metrics easily to other
> systems such as graphite, ganglia or Java's JVM (VisualVM).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
