[ 
https://issues.apache.org/jira/browse/AMQ-8414?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Matt Pavlovich updated AMQ-8414:
--------------------------------
    Description: 
Modern disks are fast, and measuring time in millis is not fine grained enough 
to get correct averages when there are a few blips in performance.

Current:
TimeStatisticImpl
{noformat}
System.currentTimeMillis() - start
{noformat}

Proposed:
NanoTimeStatisticsImpl extends TimeStatisticImpl
{noformat}
System.nanoTime() / 1000 - start
{noformat}

kahaDB config flag "statisticsUnit="nanos | millis" (default: millis)


  was:
Modern disks are fast, and measuring time in millis is not fine grained enough 
to get correct averages when there are a few blips in performance.

Current:
TimeStatisticImpl
{noformat}
System.currentTimeMillis() - start
{noformat}

Proposed:
NanoTimeStatisticsImpl extends TimeStatisticImpl
{noformat}
System.nanoTime() / 1000 - start
{noformat}




> [Proposal] Support KahaDB stats to be micro-seconds or milliseconds
> -------------------------------------------------------------------
>
>                 Key: AMQ-8414
>                 URL: https://issues.apache.org/jira/browse/AMQ-8414
>             Project: ActiveMQ
>          Issue Type: New Feature
>            Reporter: Matt Pavlovich
>            Assignee: Matt Pavlovich
>            Priority: Major
>             Fix For: 5.17.0, 5.16.4
>
>
> Modern disks are fast, and measuring time in millis is not fine grained 
> enough to get correct averages when there are a few blips in performance.
> Current:
> TimeStatisticImpl
> {noformat}
> System.currentTimeMillis() - start
> {noformat}
> Proposed:
> NanoTimeStatisticsImpl extends TimeStatisticImpl
> {noformat}
> System.nanoTime() / 1000 - start
> {noformat}
> kahaDB config flag "statisticsUnit="nanos | millis" (default: millis)



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

Reply via email to