Provisionamento automático de nós para clusters de baixo custo: uma abordagem utilizando K3s, Python e Terraform

Abstract

The increased reliance of companies and institutions on technology underscores the need for high system availability and a balance between costs and operational impacts. The IT infrastructure, encompassing the entire hardware base and operating systems on which software operates, becomes a pivotal factor. Kubernetes is an orchestrator that uses containers and virtualization to ensure replication, load balancing, portability, and ease of management, standing out for its assistance in addressing these issues. This work proposes the development of an automatic scaling algorithm for Kubernetes clusters composed of low-resource virtual machines, aiming to dynamically handle variations in resource demand, ensuring greater application availability, as well as streamlining management and reducing hardware costs. By addressing scaling issues before resource scarcity, the algorithm seeks to optimize the end-user experience and promote financial savings for companies.