Protecting Revenue at Scale: Pre-Detecting Anomalies in High Velocity ECommerce Systems using Artificial Intelligence (AI)
- Version
- Download 7
- File Size 316.36 KB
- Download
Protecting Revenue at Scale: Pre-Detecting Anomalies in High Velocity ECommerce Systems using Artificial Intelligence (AI)
Priyadarshini Jayakumar
Dennis Chan
Abstract
This paper presents a production oriented AIOps (Artificial Intelligence Operations) approach that advances conventional observability into pre-detecting anomalies early enough to prevent customer impact for large-scale digital commerce. Grounded in an Intelligent Traffic Routing program integrated with AI (Artificial Intelligence) and Automated Ops, the approach combines multi-layer health monitoring, two stage thresholding, time-aware seasonal baselines, cloud based cross stack delta localization, event aware reasoning, and confidence weighted recommendations with human in the loop, adding human approvals for high blast radius actions.
Index Terms
AIOps, anomaly pre detection, dynamic thresholds, time-series monitoring, cross stack delta analysis, incident triage, traffic routing, SRE, governance, self-healing.