Resilient DevSecOps: Self-Healing Cloud-Native Systems via SRE-Driven AI Threat Detection and Response

Cloud-native systems operate in highly dynamic environments where continuous deployment, service elasticity, and distributed execution expose platforms to evolving security threats and reliability risks. Conventional DevSecOps practices automate security integration into development pipelines; however, they often remain reactive, depend on manual intervention, and lack runtime resilience when systems experience active attacks or unexpected failures. This creates a need for architectures that can autonomously detect threats, mitigate risk, and maintain operational reliability.This paper proposes a resilience-centered DevSecOps framework that unifies Site Reliability Engineering (SRE) principles with AI-driven threat detection and automated remediation to enable self-healing cloud-native systems. The framework incorporates continuous telemetry ingestion, anomaly and threat inference, reliability-aware risk prioritization, and automated recovery mechanisms embedded into the operational lifecycle. The objective is to preserve service availability, reduce incident response latency, and contain failure impact during security events.Architectural modelling and evaluation demonstrate that the proposed approach enhances reliability posture, improves operational continuity, and supports proactive resilience under adversarial and failure-prone conditions. This work contributes an SRE-aligned DevSecOps paradigm that advances cloud-native security from reactive protection toward autonomous, reliability governed security assurance.