Advances and Challenges in AI-Driven Kubernetes Management: A Comprehensive Survey of Root Cause Analysis, Self-Healing and Observability
Year: 2026
Type: article
Abstract: The adoption of Kubernetes as a foundational container orchestration platform has revolutionized the deployment, scaling, and management of microservice-based applications. This survey provides a comprehensive review of representative studies in Kuber-netes management, anomaly detection, root cause analysis (RCA), and self-healing systems. We classify the literature into five core domains: container management and scheduling, anomaly detection and observability, AI-assisted root cause analysis, self-healing frameworks, and security-hardening mechanisms. Through critical analysis, we identify prevailing methodological approaches, including rule-based heuristics, machine learning, reinforcement learning, knowledge graph embeddings, and large language model (LLM)-assisted diagnostics. Emerging trends include proactive AI-driven monitoring, multi-agent orchestration, predictive self-healing, and explainable automated reasoning. Despite substantial advances, open challenges persist in scalability to heterogeneous and multi-cloud environments, dataset standardization, human-in-the-loop integration, and secure automated remediation. By syn-thesizing these findings, this survey provides a structured roadmap for researchers and practitioners, highlighting key gaps and opportunities to develop resilient, efficient, and intelligent Kubernetes ecosystems capable of meeting the demands of modern cloud-native and edge-aware infrastructures.
Source: DergiPark (Istanbul University)
Authors: Gundogdu, Emre, Kayaduman, Unal, Pınarer, Ozgun
Institutions: Siemens (Türkiye), Galatasaray University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204431249