Automating FinOps in Cloud Computing: An Integrated Solution for Efficient Data Collection with Dynamic Scraper Generation

This paper introduces a framework to integrate Financial Operations (FinOps) practices in Kubernetes, addressing the challenge of managing cloud services' costs across multicloud environments. The framework automates the collection of service providers' costs by deploying Prometheus exporters and scrapers, then standardizes cost data according to the FinOps Cost and Usage Specification (FOCUS) through Large Language Models (LLMs). It enables automatic data analysis, implemented by a metrics aggregator, to compute insightful cost and quality of service optimizations. We present the experimental results on three standard cloud service providers, achieving up to 91% accuracy in automatic data standardization. The results corroborate that the framework helps simplify cloud cost management and promote FinOps principles.