The Genealogy of Crisis: Unraveling Cloud Incidents from the Flow-Oriented Perspective

Cloud services have surged in popularity, offering cost-effective and elastic computing resources to global companies. Yet, cloud service providers grapple with incidents that jeopardize digital stability and security. While incident response solutions are well-documented, scant research delves into the crisis genealogy during the incubation phase. In this study, we scrutinize Google Cloud incidents from a flow-oriented perspective. Our preliminary analysis identifies four pivotal constructs during the crisis incubation period, namely the crisis origins, the flowing lines of action, the correspondence, and the triggering events. Notably, we unveil a dynamic interplay between crisis origins and triggering events—a strategic insight for incident prevention. Additionally, we identify flowing lines of action into two categories: customer-facing services degradation and company-facing systems degradation. Within these, we distinguish observational and responsive correspondences. This identification lays the groundwork for fine-tuning incident prevention strategies and bolstering service resilience in the cloud ecosystem.