Architecting Autonomous Resilience: A Closed-Loop Chaos Engineering Framework for Cloud-Native Supply Chain Logistics Platforms

Abstract

Cloud-native supply chain logistics platforms increasingly depend on distributed microservice topologies whose failure modes are difficult to anticipate through design-time analysis alone. Traditional chaos engineering practice addresses this by injecting controlled faults into production or pre-production environments, but most implementations remain a manual, human-scheduled exercise: an engineer selects a fault, injects it, and observes the outcome, separating fault discovery from remediation. This paper proposes a Closed-Loop Chaos Engineering Framework that couples automated fault injection with a continuous observability pipeline and a policy-driven remediation layer, closing the loop between failure discovery and corrective action within a single operational cycle. The framework has three stages: a fault injection controller that scopes experiments against a defined blast-radius boundary, a telemetry correlation layer that maps observed degradation back to the injected fault, and a remediation policy engine that applies pre-authorized corrective actions when resilience thresholds are breached. The paper discusses the framework's design rationale and architecture, and presents an illustrative, qualitative walkthrough of its behavior under representative logistics-platform fault scenarios rather than validated empirical results. Limitations and future empirical validation directions are discussed.