Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: jira@apache.org
Date: Wed, 8 Apr 2015 19:26:12 +0000 (UTC)
From: "Billie Rinaldi (JIRA)" <jira@apache.org>
To: notifications@accumulo.apache.org
Message-ID: <JIRA.12819345.1428505037000.30583.1428521172791@Atlassian.JIRA>
In-Reply-To: <JIRA.12819345.1428505037000@Atlassian.JIRA>
References: <JIRA.12819345.1428505037000@Atlassian.JIRA>
 <JIRA.12819345.1428505037218@arcas>
Subject: [jira] [Commented] (ACCUMULO-3715) Decrease sampling percentage for
 tracing
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/ACCUMULO-3715?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14485856#comment-14485856 ] 

Billie Rinaldi commented on ACCUMULO-3715:
------------------------------------------

My idea of allowing 0ms spans for Accumulo but not HDFS won't work because most of the HDFS spans are in the client, meaning that they'll originate in an Accumulo process.  We'd be left to distinguish HDFS spans from Accumulo spans by inspecting the span description, which sounds bad.  This makes me think we should just start dropping all 0ms spans again.  We could still make it configurable, though.

> Decrease sampling percentage for tracing
> ----------------------------------------
>
>                 Key: ACCUMULO-3715
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3715
>             Project: Accumulo
>          Issue Type: Bug
>          Components: trace
>            Reporter: Billie Rinaldi
>            Assignee: Billie Rinaldi
>            Priority: Critical
>             Fix For: 1.7.0
>
>
> Now that HDFS has tracing, individual traces might collect much more tracing data than they did previously.  We should have a way of controlling how much tracing data is collected other than just turning off HDFS tracing.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)