Too many open threads and ServiceNow integration

[Tim Deardeuff]

I have read the other topics on this and they don't apply. I am running Jira with a user account. I've verified all the files are owned by the user that Jira is running as and we aren't using NFS mounts. 

Regularly the system either runs slowly and the troubleshooter reports too many open files, or it just hangs. We have taken to restarting the system every morning just to reset the files. Clearly there is a memory leak. When I run ls -l /proc/<pid>/fd wait an hour and run it again I see that there are many more open files but the regular filesystem files are the same but many more sockets: exist.

In addition, the catalina.out log file shows dozens of messages with this as the final statement:  The web application [ROOT] appears to have started a thread named <varies> but has failed to stop it. This is very likely to create a memory leak.

I believe these statements are part of the shutdown process.

Also the suggested fixes elsewhere suggest running ulimit -aS which returns oepnfiles of 1024, even though I have /etc/security/limits.conf set to 

jira soft nofile 4096
jira hard nofile 8192
jira soft nproc 4096
jira hard nproc 8192

Does any of this give an indication of what we could have wrong? It was working fine but of course I can't pin it to any change made to the system. 

Thanks, Tim