Tony Anecito schrieb:
I have asked on this group but have never gotten an answer. My guess
is it means socket connect time. What that means is the real
question. Again, I have a theory that the amount of time is not just
the response time (taking into account the chatter of the signaling)
of the app but how long it takes before the client responds to close
the socket and the signal to propagate.
My 2cents worth. When I look at my logs on Windows the response time
as measured at Apache for requests that gets passed to Tomcat
(embedded in JBoss 4.2.2) is typically 0msec. Sometimes I see 15msec
even with -D or microsecond resolution (windows OS timer issue).
For http SOAP requests I get from all over the world from my RIA I
typically see 0msec in Apache and I am curious what is measured
elsewhere.
--- On Thu, 7/17/08, ywtsang <[EMAIL PROTECTED]> wrote:
i log the response time by the following option:
%D - Time taken to process the request, in millis
does this mean that the time includes the sum of:
- start of request - application logic - streaming the whole
response (html text) back to client
?
The time is taken at the beginning of the AccessLogValve invoke as part
of the valves chain, and at the end, directly before formatting the log
file and writing it out.
So it will not contain connection setup and I think also not receiving
the request headers. It will include everything from then until writing
the final response data to the socket.
Regards,
Rainer
---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
