http://tomcat.apache.org/tomcat-7.0-doc/config/valve.html
*%D* - Time taken to process the request, in millis
What does %D in the actually represent? Let's say the stack trace looks
like:
at org.apache.catalina.connector.CoyoteAdapter.service
(CoyoteAdapter.java:403)
at
org.apache.coyote.http11.Http11NioProcessor.process(Http11NioProcessor.java:369)
at
org.apache.coyote.http11.Http11NioProtocol$Http11ConnectionHandler.process(Http11NioProtocol.java:317)
at
org.apache.tomcat.util.net.NioEndpoint$SocketProcessor.run(NioEndpoint.java:1532)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
Does %D represent the time spent in .run(), or .process(), or .service(), or
something else?
Assume non-keepalive for the moment. What happens if a client
connects...but waits some period of time after connecting before the thread
pool can service the request? Is there a way to see (log) that amount of
time?
For example, I'm trying to tune my acceptCount. I understand the default is
100, but I'm interested in possibly cranking that up. I'd like to be able
to quantify this experiment somehow, to see how much time a client spent
waiting *after* the socket connection was established, but before the thread
pool actually serviced the request.
Any advice?
Thanks,
Dan
