[
https://issues.apache.org/jira/browse/HADOOP-16266?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16829165#comment-16829165
]
Steve Loughran commented on HADOOP-16266:
-----------------------------------------
I agree that post is out of date; I'm not sure how things have changed. Links
would be good. At the same time, in a world of containers and VMs, the base
assumptions we have of time "it moves forwards consistently" may not hold. Just
bear that in mind and consider coding for it. (I type, while thinking of those
places where I haven't)
I like the idea of using nanos; there's also all of the java 9 time APIs which
may be a bit too expensive here, especially as you aren't worrying about "real
world" dates and times, just elapsed times.
Given this is just perf metrics, go for the nanotime, use that nano suffix, but
add handling for negative values and for any per-op metrics do something to
avoid division by zero errors
> Add more fine-grained processing time metrics to the RPC layer
> --------------------------------------------------------------
>
> Key: HADOOP-16266
> URL: https://issues.apache.org/jira/browse/HADOOP-16266
> Project: Hadoop Common
> Issue Type: Improvement
> Components: ipc
> Reporter: Christopher Gregorian
> Assignee: Christopher Gregorian
> Priority: Minor
> Labels: rpc
> Attachments: HADOOP-16266.001.patch, HADOOP-16266.002.patch,
> HADOOP-16266.003.patch, HADOOP-16266.004.patch, HADOOP-16266.005.patch,
> HADOOP-16266.006.patch
>
>
> Splitting off of HDFS-14403 to track the first part: introduces more
> fine-grained measuring of how a call's processing time is split up.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
