How do you want to determine whether the service is up or not? Is it a service that answers on a TCP or UDP port? If so, it would be relatively simple to configure a health monitor to poll the server:port to see if it's listening. You can configure the LTM monitoring daemon, bigd, to send a specific string and expect a string in the response. If the service can't be checked with a standard TCP or UDP request, can you provide more detail on how you would want to verify the service is up? You could potentially write an external monitor that calls a script which performs the check from the LTM.
Aaron