Abstract: A method, apparatus, and article of manufacture for failure recovery in a computer network. The invention provides a system monitor that allows a process to register itself for monitoring and dictate a failure recovery mechanism if the process terminates abnormally. The system monitor according to the present invention continuously monitors the process and detects when the process abnormally terminates. A failure recovery mechanism is then executed. The failure recovery mechanism includes the execution of an executable, a command line script or the starting of a service. These actions allow virtually any failure recovery scenario such as re-starting the failed process, cleaning up after the failed process, notifying other processes of the failure, or sending a notification of the failure.