Sistema conversacional para monitoreo y gestión en tiempo real de servidores

Abstract

Abstract: This degree project develops and validates a conversational assistance prototype for centrally querying the status of services and resources on Linux and Windows servers. The solution integrates a locally executed language model, a conversational interface, and an API developed in Python, with tools for performing queries through SSH and WinRM. The development is organized iteratively through sprints and includes mechanisms for validation, interaction logging, evidence storage, and metric generation. The evaluation is carried out in a controlled environment through unit tests, API route tests, functional validations, and scenarios with simulated and real dependencies. The results demonstrate that the prototype enables technical queries, provides understandable responses, and maintains traceability of the operations performed. In addition, compliance with the functional requirements and most of the non-functional requirements is verified. The main limitation identified is reliability when facing connection errors, timeouts, and configuration adjustments.