I have a question about the behavior of IHS (IBM HTTP Server) when an backend application server is not responding any more (due to FullGC problems) but it is still responsive on the TCP level.
You can configure a parameter called "MaxConnections" in the IHS plugin to limit the connections an IHS establishes to an app server.
How exactly can you correlate TCP backlog of the IHS's OS with the configured value "MaxConnections" in the plugin?
Is each thread of IHS equal to a TCP connection and are threads getting into the TCP backlog counting to the parameter "MaxConnections"?