Kubernetes On-Premises deployment
Last updated on April 5, 2025
Project Overview
This project involved deploying and managing a Kubernetes cluster on-premises for a cloud distributor company. The cluster was deployed directly on three bare metal servers using kubeadm and configured for networking security and persistent storage. The setup also integrated Cilium for networking, NFS for storage, Velero for backup and migration, and Prometheus for alerting and monitoring.
The Challenge
This cloud distributor company was a newly established business, having been in operation for less than one year when we undertook this project. They managed multiple websites—some for internal operations and others for external purposes, such as landing pages and cloud storage services.
Their DevOps team was small, consisting of only two people: one handling infrastructure and the other responsible for development. Due to these constraints, they needed an infrastructure that was developer-friendly, scalable, and easy to manage.
From a developer’s perspective, traditional virtual machines were complex and resource-intensive, making them less suitable for rapid application development. Instead, they opted to focus on containers, which provided greater flexibility. Containers could run on virtual machines, bare metal servers, and public cloud environments, as well as within their own data center. Given this, Kubernetes became the ideal choice for orchestrating their containerized applications.
Project Objectives
Deployed Kubernetes Clusters on three nodes.
Project deliverables were mainly the following:
- Deploy a scalable and resilient Kubernetes cluster
- Ensure secure networking with Cilium
- Enable persistent storage using NFS
- Implement a reliable backup and disaster recovery strategy with Velero
- Facilitate secure external access to services
Result
As a result, we managed to deploy a fully operational Kubernetes cluster on bare metal, providing the company with a scalable and efficient infrastructure.
The company has benefited from greater portability since Kubernetes provides a consistent API; they can move their web application very easily between namespaces for experimental purposes.
Being able to use their own servers when appropriate is lowering costs and providing them access to hardware that they wouldn't necessarily have access to in the cloud. As long as the utilization is high, the costs are much lower.
Launching a development and staging environment in different namespaces also takes far less time.
Overall, the new infrastructure enabled the company to run its workloads more efficiently, securely, and with greater flexibility, all while being developer-friendly and easy to maintain.
Download PDF