[ 
https://issues.apache.org/jira/browse/IMPALA-14878?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18074746#comment-18074746
 ] 

ASF subversion and git services commented on IMPALA-14878:
----------------------------------------------------------

Commit 7ea5c35245b49629c6a35b38041ff0144d6dd9a1 in impala's branch 
refs/heads/master from Joe McDonnell
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=7ea5c3524 ]

IMPALA-14878: Keep incoming_queue_time separately for control and datastream 
services

Incoming queue time measures the time between the RPC reactor thread
dispatching something to the service queue and a service thread picking
it up. Currently, the control service and datastream service use the
same histogram for incoming_queue_time. Since the incoming_queue_time
is impacted by the availability of service threads, using separate
histograms for the different services gives better information about
the response times.

This uses separate histograms for the different services. Beyond the
control service and datastream service, this also required changing
admission control and unit tests.

Testing:
 - Hand tested to verify that the histograms are being maintained
   separately

Change-Id: I89e5dad9e144a7b8506c70e23c043b9d0b3142dc
Reviewed-on: http://gerrit.cloudera.org:8080/24180
Reviewed-by: Michael Smith <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
Reviewed-by: Yida Wu <[email protected]>


> Use separate incoming_queue_time histograms for control service vs data 
> stream service
> --------------------------------------------------------------------------------------
>
>                 Key: IMPALA-14878
>                 URL: https://issues.apache.org/jira/browse/IMPALA-14878
>             Project: IMPALA
>          Issue Type: Bug
>          Components: Distributed Exec
>    Affects Versions: Impala 5.0.0
>            Reporter: Joe McDonnell
>            Priority: Major
>
> The "incoming_queue_time" histogram on the RPCZ page is shared between the 
> control service and the data stream service. It would be useful to have a 
> separate histograms for the two, as they use separate service threads and can 
> have different queuing times.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to