An Empirical Analysis of Server Utilization and Optimization Strategies in a Proxmox VE Environment

Server underutilization in enterprise and institutional data centers represents a significant source of financial waste and environmental impact. While proprietary virtualization platforms are extensively documented in literature, open-source alternatives like Proxmox Virtual Environment (VE) require further empirical analysis to validate their role in cost-effective IT management. This paper presents a detailed case study investigating the resource utilization of two heterogeneous Proxmox VE servers deployed in a corporate data center. Through a multi-resolution analysis (yearly, monthly, weekly, daily, hourly) of telemetry data – including CPU, memory, swap, load average, and network statistics – we diagnose critical inefficiencies: severe memory over-commitment on one server and systemic underutilization on the other. These patterns highlight common pitfalls such as imbalanced resource allocation, inadequate workload distribution, and insufficient VM right-sizing. Based on our findings, we propose a set of actionable optimization strategies, including cross-server workload migration, dynamic resource allocation, and energy-aware consolidation. Our results, derived from a real-world deployment in one of the biggest technopolis in Türkiye, demonstrate that a data-driven approach to managing Proxmox VE infrastructures can substantially reduce capital and operational expenditures, mitigating the rising costs of server hardware and operations.