Advanced Kubernetes provisioning, scaling and operations through kOps library

Kubernetes has emerged as the de facto standard for deploying and managing applications that scale from a handful of users to millions, becoming a cornerstone of the modern cloud-native ecosystem. It makes highly scalable and resilient infrastructures possible when paired with cloud computing. However, Kubernetes adoption remains hindered by two major obstacles: a steep learning curve, particularly for small or non-specialized teams, and the persistent issue of vendor lock-in within cloud platforms. Vendor-specific technologies, proprietary APIs, and economic constraints often limit portability and complicate multi-cloud strategies, creating barriers to innovation and flexibility. This thesis looks into the difficulties of Kubernetes adoption and explores solutions for simplifying cluster provisioning and management. An extensive review of the state of the art was conducted, categorizing and comparing the most widely used provisioning approaches, including self-managed installers such as Kubeadm, automation frameworks like Kubespray, Infrastructure-as-Code (IaC) tools such as Terraform and Pulumi, and managed services offered by major providers, including Amazon EKS, Google GKE, and Azure AKS. Each approach was evaluated in terms of complexity, supported environments, and trade-offs, with special focus on open-source alternatives that promote transparency and flexibility. The core contribution of this work is the development of an Elemento–kOps plug-in that extends kOps, an open-source library widely adopted for provisioning production-grade Kubernetes clusters, for the Elemento cloud platform. Elemento, an Italian deep-tech provider, distinguishes itself through its AtomOS hypervisor, which enables hybrid and multi-cloud deployments. By introducing this plug-in, the project delivers a unified interface for cluster lifecycle management, offering an alternative that combines portability, simplicity, and scalability. Methodologically, the research involved analysing the kOps codebase, adopting debugging and logging strategies to navigate its architecture, and leveraging test-driven exploration of APIs and provider-specific implementations. The integration was realized through a translation layer between kOps and the Elemento Go library, ensuring seamless provisioning of virtual machines, storage, and networking resources through declarative tasks. Evaluation was performed by comparing the provisioning speed, reliability, and code complexity of the Elemento integration against conventional cloud providers. The results demonstrate the technical feasibility and potential advantages of supporting Elemento within kOps. The integration not only simplifies the provisioning process but also enables access to a distributed network of smaller cloud providers, contributing to reduced vendor lock-in and fostering a more open, flexible ecosystem. At the same time, limitations were observed in terms of reliability testing and the need for broader validation across diverse environments. In conclusion, this work highlights the importance of open-source tools in shaping the future of cloud-native infrastructure. The Elemento–kOps integration illustrates how collaborative development can expand the accessibility of Kubernetes while mitigating structural challenges in the cloud landscape. Future improvements may include enhancing system autonomy, expanding provider interoperability, and contributing further to the open-source community to support sustainable, vendor-neutral cloud adoption.