Mailing-List: contact beehive-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Beehive Developers" <beehive-dev@incubator.apache.org>
Message-ID: <67715981.1118780628315.JavaMail.jira@ajax.apache.org>
Date: Tue, 14 Jun 2005 22:23:48 +0200 (CEST)
From: "David Read (JIRA)" <beehive-dev@incubator.apache.org>
To: beehive-dev@incubator.apache.org
Subject: [jira] Created: (BEEHIVE-814) Provide ability to track lifetime of
 PageFlow instances
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Provide ability to track lifetime of PageFlow instances
-------------------------------------------------------

         Key: BEEHIVE-814
         URL: http://issues.apache.org/jira/browse/BEEHIVE-814
     Project: Beehive
        Type: New Feature
  Components: NetUI  
    Versions: TBD    
 Environment: All
    Reporter: David Read
 Assigned to: Rich Feit 
    Priority: Minor
     Fix For: TBD


We've been looking at how to capture the lifetime of PageFlow instances through the EventReporter API.  The original thinking was to generate create and destroy events that were correlated with an instance identifier for the PageFlow.  Someone who wanted to analyze PageFlow usage could then "join" the create/destroy events to understand (1) how many of what type of PageFlow were created (over what period of time) and (2) how long PageFlows of each type existed.

We originally looked at using system identity hash code to identify an instance (session id doesn't work since the same flow can exist more than once in a given session ... at the same time or over time).  The problem with that approach is when clustering is involved.  If the destroy event happened on a "secondary" server, the (calculated) correlation value would be different.

Two other approaches come to mind ... but there could be others ...

(1) capture the create time as a non-transient attribute of the flow and expose that via a public API.  That would allow the event reporter to ignore the create event (WRT lifetime), and just capture the destroy event.  While this does add to memory and serialization cost, the benefit is that it is fairly small and is fixed.

(2) capture an identity as a non-transient attribute of the flow and expose that via a public API.  This may be a more useful concept, but might have a larger (or more variable) impact on performance. 


-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira