A Serverless Event-Driven Architecture for Automated Customer Feedback Management Using AWS SNS, Lambda, and CodePipeline

Abstract: Organizations increasingly depend on timely customer feedback to refine services, yet conventional feedback- handling pipelines are often built on continuously running servers that are costly to operate, slow to react to demand surges, and burdensome to maintain. This paper presents the design and empirical evaluation of a fully serverless, event- driven platform that automates the ingestion, analysis, routing, and escalation of customer feedback. The architecture is realized on Amazon Web Services, where submissions enter through an API gateway, propagate as events through a publish-subscribe notification service, and are processed by stateless functions written in Python that perform validation, sentiment classification, and category-based routing. A lightweight web client built with Node.js provides submission and administrative interfaces, while processed records are persisted in a managed NoSQL store and surfaced through an analytics dashboard. The entire delivery lifecycle build, test, and deployment is automated through a managed continuous-integration and continuous-delivery pipeline governed by Infrastructure-as-Code. Experimental evaluation under synthetic event load shows that the platform sustains an end-to-end latency of approximately 210 ms at 1000 events per second, scaling elastically where a server-based baseline degrades beyond 1400 ms. Pipeline automation reduced deployment lead time from roughly 75 minutes to under 10 and lowered processing cost per million requests by nearly sixfold owing to consumption-based billing. The contributions comprise an integrated serverless reference architecture for feedback automation, a reproducible CI/CD strategy, and a quantitative comparison establishing the efficiency of event-driven designs for this class of workload.

Keywords: Serverless computing; Event-driven architecture; AWS Lambda; Amazon SNS; Continuous delivery; Sentiment analysis; Feedback management; Cloud automation