Hello all, Since yesterday all of our loadbalanced tomcat servers have been refusing further connection intermittently. During our internal stress testing, we’ve managed to reproduce the issue where maximum number of threads/connections to Catalina were reached for one server, but then the rest were loadbalanced to others. Eventually, everyone starts failing health checks the Citrix network health monitor on port 80 and 8443. Because all of them were eventually running out.
We wanted to pinpoint the time of the first hit observers with Catalina logs to see when max connections are reached or maxThreads are used. Is there any log level for Catalina/localhost logs which we can use to see this message ? If so which level should we be using ? Also, we are using 8.5.37 with our own TSL cipher suites etc. this started happening only since last few days. We would need to find out whether any application threads are doubling up and exhausting the APR connectors capacity for max threads/connections by holding them indefinitely. Any suggestion/help you can give here would be appreciated. Thanks,
