Instance B knows nothing about instance A. Instance B is only serving responses to instance A. Instance A might do a lot of work before and after it talks to instance B which might make instance B look bad even if that is the short part of a transaction.

The only gotcha is if client side javascript were introduced. The javascript might be doing calls back to A introducing a nasty piece to the puzzle.

-Tim

Jeremy Nix wrote:

Okay, let me pose another (slightly differenc) question.  Say that
instance (A) and (B) are separate institutions, independent from each
other.  Same scenario as before.  How could instance (B) (the responding
instance) be able to measure latency in instance (A)?  The reason I ask
is related to an issue that has been reported to me, but I am unable to
understand how the latency was measured.


_________________________________ Jeremy Nix Southwest Financial Services, LTD. [EMAIL PROTECTED] (513) 621-6699 ext.1158



-----Original Message-----
From: Tim Funk [mailto:[EMAIL PROTECTED] Sent: Wednesday, January 21, 2004 8:26 AM
To: Tomcat Users List
Subject: Re: Measuring latency between tomcat instances



LOG4j might be nice kludge for this. On tomcat A, time the appropriate
events on how long it takes to do stuff on tomcat B. Then log it in a
specially formatted message. (mylogger.log("time: " + timeTaken) Then do the same
on tomcat B. Its easier to do on tomcatB since you can use a Valve or
servlet filter and stay out of the business logic.


I say LOG4J since it provides asynchronous loggers (Unless the
performance on the non-asynchronous loggers are neglible in this case). But with LOG4j
(or insert logging framework here) - you can use it to write to any log file
or format you want. So your normal error messages can go to one file while
these special messages can go to a differnet file/database/...


------

Or you can compare acces log entries since it does offer a %T option for
time taken.


-Tim

Jeremy Nix wrote:


We have an instance where tomcat instance (A) is communicating with tomcat instance (B), where instance (B) is located at another site. Instance (A) formulates a request to instance (B) and instance (B) responds with some valid response. Now, my question is, from a standpoint of instance (B), is it possible to measure latency? If so,


how?




--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]


--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]





---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to