Agentic AI Framework for Autonomous and Self-Managing Cloud Services

The growth in cloud-native applications has made their usage hard to plan and monitor. The rule-based Kubernete execution is more active and flexible than its peers, which in itself is a challenge when the cloud is evolving in obscure ways. This is the reason behind our framework which utilizes multiple agent deep learning for needs application orchestration. It is an orchestration framework that employs Multi-Agent Reinforcement Learning (MARL) to permit crafted agents to optimize the method and goals simultaneously and adapt to variable parameters. We explain the components of the structure and deep learning models that focus on decision making in the framework architecture. Using a cloud cluster in a simulation, we were able to show that our approach, unlike the other orchestration systems, did greatly better in relation to resource usage, service latency, and operational resilience. This is strong evidence for the intelligent, self-managing cloud architecture we propose.