[
https://issues.apache.org/jira/browse/CASSANDRA-11569?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15293227#comment-15293227
]
stone edited comment on CASSANDRA-11569 at 5/20/16 11:58 AM:
-------------------------------------------------------------
I understand the latency calculate method now.what i still confuse is
crossNodeLatency should be calculated when a message is writed,not calculate as
time goes by.
for example,I have no workload,the metrics count should not increase,and the
mean should keep same as no new data.
but actually when the crossDCLatency is large,about 45ms.the count never
increase.I know there may caused by misunderstanding node status.and I need to
increase the connect tries to check the crossDC node,which may do an impact on
performance
was (Author: stone):
I understand the latency calculate method now.what i still confuse is
crossNodeLatency should be calculated when a message is writed,not calculate as
time goes by.
for example,I have no workload,the metrics count should not increase,and the
mean should keep same as no new data.
but actually when the crossDCLatency is large,about 45ms.the count never
increase.
> Track message latency across DCs
> --------------------------------
>
> Key: CASSANDRA-11569
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11569
> Project: Cassandra
> Issue Type: Improvement
> Components: Observability
> Reporter: Chris Lohfink
> Assignee: Chris Lohfink
> Priority: Minor
> Attachments: CASSANDRA-11569.patch, CASSANDRA-11569v2.txt,
> nodeLatency.PNG
>
>
> Since we have the timestamp a message is created and when arrives, we can get
> an approximate time it took relatively easy and would remove necessity for
> more complex hacks to determine latency between DCs.
> Although is not going to be very meaningful when ntp is not setup, it is
> pretty common to have NTP setup and even with clock drift nothing is really
> hurt except the metric becoming whacky.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
