The message is generic and seen often for Essbase servers installed on Linux/Unix server.Unless you see a performance issue with starting applications or any other action you can ignore the message.
You would need to check what caused the 32K limit to be reached. You could increase the soft limit value, until the hard limit by changing the limit.conf file or its snippets.
If you don't some of the files wouldn't open or you would have a jvm error. There might be other errors too.
I have also noticed that the "the 32K limit has been exceeded" message does not change even if i keep different values for the soft limit like below.
hypXXX soft nofile 32000
hypXXX soft nofile 16000
hypXXX soft nofile 8192
Which means the message is not depended on the value i set here?
Are you setting the nofile soft limit as a root user or a normal user? You can set the limits for an individual user or for all.
You would need to reboot your server for the settings to take effect.
One of the causes for this might be that there are too many requests waiting instead of completing sooner. Some of the effects might be that applications wouldn't load, essbase might not start, or users might not be able to login. It would be an OS issue but would interfere with the working of hyperion.
Can you run the following on the server?
netstat -a | grep TIME_WAIT | wc -l
Also, you could change your timeout setting so that requests time out and resources are freed.