Generative AI for Cloud Infrastructure Decision-Making and Self Healing Systems

Abstract

Cloud infrastructure has grown increasingly complex, demanding intelligent automation to ensure performance, reliability, and resilience. This paper explores the application of Generative Artificial Intelligence (Generative AI) to enhance decision-making and enable self-healing capabilities in cloud environments. Generative models such as large language models (LLMs), generative adversarial networks (GANs), and variational autoencoders (VAEs) are proving instrumental in addressing challenges related to dynamic resource provisioning, anomaly detection, root cause analysis, and automated remediation. I present a framework that leverages generative models to simulate failure scenarios, generate configuration policies, and synthesize runbooks for autonomous recovery. Integration with observability pipelines and cloud-native services enables closed-loop, real-time adaptation, reducing mean time to resolution (MTTR) and improving system uptime. Case studies demonstrate improved accuracy in fault prediction and faster recovery compared to traditional methods. I also discuss implementation challenges, including model drift, latency constraints, and data privacy. This study underscores the transformative potential of Generative AI in building resilient, adaptive, and scalable cloud infrastructures, while offering practical insights for architects,DevOps teams, and AI researchers aiming to advance autonomous cloud operations.