Researchers have been emphasizing the strong role that multi-agent systems can play progressively towards the design and implementation of complex autonomic systems. The aim of this research has been to provide the agent programmers with a very model that allows the design and construction of entities or agents possessing the desired autonomic behaviours primarily including goal-driven self-adaptation, self-healing and self-optimization developed upon a reactive, behaviour based architecture. This paper introduces the conceptualization and the design of the autonomic agent architecture for the Scalable fault tolerant Agent Grooming Environment (SAGE) – a second generation FIPA-Compliant multi-agent system.