A Study on CI/CD and Deployment Strategies: A Constructive Study on DevOps Transformation in Microservice-based IoT SaaS Environments
Niketeghad, Erfan (2025)
2025
Tietojenkäsittelyopin maisteriohjelma - Master's Programme in Computer Science
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2025-12-02
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-2025120211181
https://urn.fi/URN:NBN:fi:tuni-2025120211181
Tiivistelmä
Internet of Things (IoT) and Software as a Service (SaaS) companies face unique challenges in implementing effective DevOps practices due to their complex microservice architectures and single-tenant deployment models. This thesis addresses these challenges through a constructive research approach, implementing and evaluating DevOps transformation practices at Treon, a company specializing in microservice-based IoT SaaS solutions. The research focused on three primary objectives: implementing a continuous deployment solution using GitOps principles, evaluating deployment strategies suitable for single-tenant IoT architectures, and standardizing continuous integration pipelines across multiple repositories. The study employed ArgoCD as the GitOps tool to automate deployment processes and enhance configuration management, systematically evaluated canary, blue-green, and rolling update deployment strategies, and enhanced existing Jenkins shared libraries to reduce code duplication and improve pipeline reliability.
Key findings demonstrate that GitOps implementation significantly improved deployment reliability, visibility, and operational workflows while reducing manual intervention and human error. The evaluation revealed that namespace-level rolling updates proved most suitable for single-tenant IoT architectures. Canary deployments were incompatible with isolated client environments, while blue-green deployments presented prohibitive resource overhead. CI pipeline standardization through enhanced shared libraries successfully reduced code duplication and improved maintainability across the microservice architecture.
The implementation resulted in observable improvements in team workflows, including enhanced deployment reliability, automatic configuration drift detection, improved team accessibility to deployment processes, and streamlined operational procedures. The research contributes practical solutions for IoT DevOps transformation and theoretical insights into adapting standard DevOps practices for single-tenant IoT SaaS environments, providing actionable guidance for companies facing similar architectural constraints.
Key findings demonstrate that GitOps implementation significantly improved deployment reliability, visibility, and operational workflows while reducing manual intervention and human error. The evaluation revealed that namespace-level rolling updates proved most suitable for single-tenant IoT architectures. Canary deployments were incompatible with isolated client environments, while blue-green deployments presented prohibitive resource overhead. CI pipeline standardization through enhanced shared libraries successfully reduced code duplication and improved maintainability across the microservice architecture.
The implementation resulted in observable improvements in team workflows, including enhanced deployment reliability, automatic configuration drift detection, improved team accessibility to deployment processes, and streamlined operational procedures. The research contributes practical solutions for IoT DevOps transformation and theoretical insights into adapting standard DevOps practices for single-tenant IoT SaaS environments, providing actionable guidance for companies facing similar architectural constraints.