[
https://issues.apache.org/jira/browse/HBASE-15163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15116755#comment-15116755
]
Elliott Clark commented on HBASE-15163:
---------------------------------------
bq.This is what you mean 'continuous couting', right?
It always goes up. Never down. So if whatever collects the metrics from HBase
isn't synced up with what's the Hadoop Metrics2 snapshot time, nothing is lost
but some resolution.
Anything that's a rate can miss lots of data and give very misleading values if
the times aren't synced up.
Say you scrape data from HBase once 5 mins. However the metrics system takes a
snapshot once every min. Then there are four periods that are completely hidden
from your dashboards.
> Add sampling code and metrics for get/scan/multi/mutate count separately
> ------------------------------------------------------------------------
>
> Key: HBASE-15163
> URL: https://issues.apache.org/jira/browse/HBASE-15163
> Project: HBase
> Issue Type: Sub-task
> Reporter: Yu Li
> Assignee: Yu Li
> Attachments: DifferentRequestQPS.png, HBASE-15163.patch
>
>
> This way we could see QPS of different kinds of requests, to better analyze
> what's causing hot spot in system
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
