polito.it
Politecnico di Torino (logo)

Practical application of Agile methodology, DevOps automation and Cloud Native Architecture principles to Data API services

Qiyang Deng

Practical application of Agile methodology, DevOps automation and Cloud Native Architecture principles to Data API services.

Rel. Antonio Vetro', Marco Torchiano. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2025

[img]
Preview
PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB) | Preview
Abstract:

Practical application of Agile methodology, DevOps automation and Cloud Native Architecture principles to Data API services Modern cloud-native systems demand scalable, self-healing architectures to manage multi-tenant environments efficiently. This paper presents the design and implementation of an autonomous tenant management framework for MZinga.io, a cloud-native data API service. The solution integrates Kubernetes Operators, GitOps workflows (leveraging ArgoCD), and an event-driven architecture grounded in RabbitMQ. These components collectively mitigate the shortcomings of conventional Azure DevOps pipeline-centric methodologies, which often introduce external dependencies and complicate on-premise deployment scenarios. The principal contributions of this work include: Declarative Resource Orchestration: Custom Resource Definitions (CRDs) streamline the management of tenants, projects, and environments, facilitating cluster-native resource orchestration without external tooling. GitOps-Driven Automation: ArgoCD synchronizes Kubernetes configurations with version-controlled Git repositories, resulting in a 98% success rate for deployments and a 40% reduction in manual operational tasks. Resilient Event Handling: A hybrid messaging system replaces webhooks with RabbitMQ, achieving 99.5% message delivery reliability while maintaining compatibility with heterogeneous cloud and on-premise environments. Integrated Observability: A unified monitoring stack combining Prometheus and Grafana delivers real-time visibility into RabbitMQ clusters and Kubernetes resources, complemented by configurable alerting mechanisms for proactive incident management. Experimental validation demonstrates a 35% reduction in operational costs and a 60% improvement in deployment speed compared to traditional methods. The framework’s modular design ensures compatibility with hybrid cloud environments, making it a robust solution for enterprises seeking agility and scalability in API service management.

Relatori: Antonio Vetro', Marco Torchiano
Anno accademico: 2024/25
Tipo di pubblicazione: Elettronica
Numero di pagine: 88
Soggetti:
Corso di laurea: Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering)
Classe di laurea: Nuovo ordinamento > Laurea magistrale > LM-32 - INGEGNERIA INFORMATICA
Aziende collaboratrici: Newesis Srl
URI: http://webthesis.biblio.polito.it/id/eprint/35263
Modifica (riservato agli operatori) Modifica (riservato agli operatori)