I'm having an app where a single request can take >30 secs. When running the app in CirrAS I'm getting 503-errors from the front-end appliance (which looks like a timeout). How can I configure the CirrAS cluster to avoid such a timeout?
This is of course not an valid behaviour. Take a look at httpd error log (/var/log/httpd/). If there is nothing unusual, see back-end JBoss AS logs.