Which of the following metrics is commonly monitored in a data center for capacity planning?

Effective data center capacity planning requires a comprehensive approach that monitors all critical infrastructure pillars: compute, storage, and network. Memory utilization is a key metric for compute resources, indicating how much server memory is in use. Disk I/O (Input/Output Operations Per Second) is a fundamental performance metric for storage systems, crucial for understanding workload demands. Network latency measures the delay in data transmission, which is vital for application responsiveness. Monitoring all these metrics together provides a holistic view, enabling administrators to proactively identify potential bottlenecks and plan for future growth across the entire data center stack.

A. Memory utilization: This is incorrect because it represents only the compute aspect of capacity planning, ignoring the equally critical storage and network components.
B. Disk I/O: This is incorrect as it focuses solely on storage performance capacity, neglecting essential compute (memory) and network (latency) metrics.
D. Network latency: This is incorrect because it addresses only the network performance dimension, while a complete capacity plan must also include compute and storage resources.

1. Official Vendor Documentation: HPE InfoSight for servers constantly monitors and analyzes server resources to provide insights for capacity planning. The collected data includes "server resources (processor, memory, I/O) and server configuration." This confirms that memory and I/O are key metrics.

Source: HPE InfoSight for servers User Guide, Edition: 11 (April 2023), Page 7, Section: "Data that HPE InfoSight for servers collects".

2. Official Vendor Documentation: HPE OneView is a foundational tool for infrastructure management and provides the data necessary for capacity planning. It monitors the health and utilization of server hardware subsystems, including processors, memory, I/O, and networking.

Source: HPE OneView 8.3 User Guide, (December 2023), Page 133, Section: "Server hardware", "Health status".

3. Academic Publication: In academic literature on data center management, capacity planning is defined by monitoring and forecasting resource usage. Key performance indicators (KPIs) include CPU utilization, memory usage, disk I/O rates, and network throughput/latency, as these directly impact application performance and scalability.

Source: Barroso, L. A., Hölzle, U., & Ranganathan, P. (2018). The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines (3rd ed.). Morgan & Claypool. Chapter 2, "Workloads and Software Infrastructure," discusses the resource demands of applications, implicitly covering memory, I/O, and network performance as planning factors.