Elisa Bariselli
Intent‑Driven Kubernetes Manifest Generation: An Agentic Approach for Cloud‑Native Workloads.
Rel. Alessio Sacco, Guido Marchetto. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2026
|
Preview |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (11MB) | Preview |
Abstract
Cloud-native systems have revolutionized the development, deployment, and management of modern applications, enabling organizations to achieve unprecedented scalability and resilience. Despite these advantages, configuring and operating such environments remains complex, particularly for users without extensive technical expertise. A recent industry trend, Agentic AI, leverages large language models (LLMs) capable of orchestrating multiple tools to automate configuration tasks and reduce this operational burden. This thesis presents an intent‑based framework that employs LLMs to automate the generation of configurations in cloud‑native ecosystems, using Kubernetes workloads as the primary case study. Although LLMs have achieved remarkable progress in natural language generation, producing correct, reliable, and domain‑specific structured configurations remains challenging.
To address this, we design a complete pipeline that begins with intent interpretation and prompt formulation, proceeds through iterative manifest generation, and incorporates validation mechanisms and error‑aware refinement loops to progressively improve configuration quality across varying levels of complexity
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
URI
![]() |
Modifica (riservato agli operatori) |
