You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by "Srimanth Gunturi (JIRA)" <ji...@apache.org> on 2013/08/23 23:25:51 UTC

[jira] [Updated] (AMBARI-3017) Service metrics call latency increases by a factor of 2 or 3 every 20-30 calls.

     [ https://issues.apache.org/jira/browse/AMBARI-3017?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Srimanth Gunturi updated AMBARI-3017:
-------------------------------------

    Summary: Service metrics call latency increases by a factor of 2 or 3 every 20-30 calls.  (was: UI's service call latency increases by a factor of 2 or 3 every 20-30 calls.)

The measurements were taken by Apache JMeter software package on a 4-node local VM cluster. JMeter which can output a 'Response Time' graph which includes Latency time (time it took for server to start responding), and Download time (byte transfer). The Download time was insignificant comapred to latency, hence the picture holds true for latency issues.

It is worth noting how JMeter generates graphs. If a thread made 500 calls, it does not show 500 points and connect them. It samples across a time period and plots the calculated time at that sample. 

So if your sampling interval is 500ms, and the run took 5 minutes - instead of 500 points, we will see 600 points. But the general trends are accurately shown.
                
> Service metrics call latency increases by a factor of 2 or 3 every 20-30 calls.
> -------------------------------------------------------------------------------
>
>                 Key: AMBARI-3017
>                 URL: https://issues.apache.org/jira/browse/AMBARI-3017
>             Project: Ambari
>          Issue Type: Bug
>          Components: client
>    Affects Versions: 1.4.0
>            Reporter: Srimanth Gunturi
>            Assignee: Mahadev konar
>              Labels: performance
>             Fix For: 1.4.0
>
>         Attachments: servicesGC.png
>
>
> Every 20-30 calls for service metrics, latency of the next 5 calls are a multiple of 2 or 3. This might be due to garbage collection. But basically in UI, the updates seem to take longer. 
> The call UI makes is
> {noformat}
> /api/v1/clusters/${cluster}/services?fields=components/ServiceComponentInfo,components/host_components,components/host_components/HostRoles,components/host_components/metrics/jvm/memHeapUsedM,components/host_components/metrics/jvm/memHeapCommittedM,components/host_components/metrics/mapred/jobtracker/trackers_decommissioned,components/host_components/metrics/cpu/cpu_wio,components/host_components/metrics/rpc/RpcQueueTime_avg_time,components/host_components/metrics/flume/flume,components/host_components/metrics/yarn/Queue
> {noformat}

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira