I do the following (this is for all our batch jobs):
- the wrapper script for SAS bacth jobs determines a timestamp and process number
- timestamp and process number are used to create the filename for the log file, used with the -log commandline option
- the wrapper script runs a quick SAS program that writes all pertinent information (parameters, program name, time etc) to a SAS dataset; return code is set to -1, and a HTML navigation page for the current day is written, containing all jobs from the day, with links to the log file location
- once the main job has run, a similar SAS program runs that adds the final return code to the jobcontrol dataset, and recreates the navigation page
As soon as a job runs, I get an entry in today's jobcontrol page on the web server (return code is "running"). Clicking on the link that is attached to the program name will display the log. With "reload" I can watch the job's progress.