flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-1792) Improve TM Monitoring: CPU utilization, hide graphs by default and show summary only
Date Fri, 10 Apr 2015 13:08:12 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1792?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14489570#comment-14489570
] 

ASF GitHub Bot commented on FLINK-1792:
---------------------------------------

Github user rmetzger commented on a diff in the pull request:

    https://github.com/apache/flink/pull/553#discussion_r28142528
  
    --- Diff: flink-runtime/src/main/scala/org/apache/flink/runtime/taskmanager/TaskManager.scala
---
    @@ -129,6 +129,25 @@ class TaskManager(val connectionInfo: InstanceConnectionInfo,
         override def getValue: Double =
           ManagementFactory.getOperatingSystemMXBean().getSystemLoadAverage()
       })
    +  metricRegistry.register("cpuLoad", new Gauge[Double] {
    +    override def getValue: Double = {
    +      try{
    +        val osMXBean = ManagementFactory.getOperatingSystemMXBean().
    +          asInstanceOf[com.sun.management.OperatingSystemMXBean]
    +        return fetchCPULoad(osMXBean).asInstanceOf[Double]
    +      } catch {
    --- End diff --
    
    On JVMs not having the `OperatingSystemMXBean` we'll log a full stack trace on each heartbeat
(each time we collect metrics).
    
    I would only register the "cpuLoad" metric if we know that we can cast the OSMXBean to
`OperatingSystemMXBean` and when the `getProcessCpuLoad()` method is available.
    Otherwise, I would register a dummy method that is always returning -1


> Improve TM Monitoring: CPU utilization, hide graphs by default and show summary only
> ------------------------------------------------------------------------------------
>
>                 Key: FLINK-1792
>                 URL: https://issues.apache.org/jira/browse/FLINK-1792
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Webfrontend
>    Affects Versions: 0.9
>            Reporter: Robert Metzger
>            Assignee: Sachin Bhat
>
> As per https://github.com/apache/flink/pull/421 from FLINK-1501, there are some enhancements
to the current monitoring required
> - Get the CPU utilization in % from each TaskManager process
> - Remove the metrics graph from the overview and only show the current stats as numbers
(cpu load, heap utilization) and add a button to enable the detailed graph.



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

Mime
View raw message