I am trying to find out the processes on our linux servers that are responsible for load on the system. We have an application server and an Oracle database server (both running Linux).
The application server runs processes which are mostly shell scripts, perl scripts and standard applications like Apache. Many of the processes open sessions in database.
I am sampling the load average on both servers once a minute using 'uptime' command.
I am also collecting the output of 'ps' command once a minute, to list the processes running.
Given the above information, what algorithms or machine learning techniques can I use to classify processes into 10 different buckets with 10 indicating the processes that cause heaviest load on the system and 1 the lightest load?