Abstract

A method and apparatus are disclosed for monitoring and managing the applications and resources on a distributed computer network. Preferably, at least one manager software system runs on at least one of the networked computer systems designated as a network management computer system or "console" system. An agent software system runs on each of the server computer systems in the network to be monitored. Each respective agent software system carries out tasks on the computer system in which it is installed such as discovering which resources and applications are present on the computer system, monitoring particular aspects of the resources and applications present on the computer system, and executing recovery actions automatically when such actions are warranted. The agents are capable of intelligent, autonomous operation. Knowledge modules are stored in a non-volatile storage device at the site of each agent software system and are loaded and unloaded into server memory dynamically as consoles register and de-register with the agents. Consoles may register to receive all information from the agents or only selected information. An event management procedure is disclosed for coordinating event management between the various consoles throughout the network.