"vola" wrote : Because I am working on self-adaptive system. If too many exceptions are thrown, I hope the system can self-adapt to solve the problem by itself. | Theoritically there can be couple of ways which comes to my mind , but don't do it unless there is not workaround 1) Modify the InternalManagedConnectionPool , have the variable in the Counter which will should be updated here . else | { | // UPDATE THE COUNTER HERE | // we timed out | throw new ResourceException("No ManagedConnections available within configured blocking timeout ( " | + poolParams.blockingTimeout + " [ms] )"); | } In the InternalManagedConnectionPool$Counter keep getFailedCount()/setFailedCount methods , these should again be exposed through the JbossManagedConnectionPool . This is trivial change in http://anonsvn.jboss.org/repos/jbossas/trunk/connector/src/main/org/jboss... http://anonsvn.jboss.org/repos/jbossas/trunk/connector/src/main/org/jboss... 2) Have a Mbean which will scan the server.log file and check for the occurance of the NoManagedConnection Error . Or you can direct the error logging to a specific log file , the Mbean will use this log file and change the setting of the Pool on the fly . But DON't try these options , you should identify the peak load to which pool can be subjected and configure max pool size accordingly . The pool will shrink if these connections are not used over long time which is configurable (idle-timeout-minutes) , so keeping the high value for the pool will not be the issue , the pool implementation will take care of all . View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4131596#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...