A Windows Server 2022 file server uses a six-disk RAID-10 array and is monitored at 30-second intervals. During a time when users complain of slow file saves, the collected averages are:
CPU utilization: 42 %
Available memory: 6 GB of 16 GB
Network throughput: 150 Mb/s on a 1 Gb/s NIC
Average Disk Queue Length: 14
Which server subsystem is most likely the performance bottleneck?
The disk subsystem is the likely culprit. An Average Disk Queue Length greater than roughly 2 outstanding requests per physical spindle indicates the disks cannot service I/O fast enough. With six disks, the recommended threshold would be about 12 (6 × 2). A measured value of 14 exceeds that limit, showing the array is saturated. By contrast, CPU utilization is well below the common 80 % alert level, ample RAM remains free, and network utilisation is only about 15 % of link capacity-none of which point to contention. Therefore, slow file saves are best explained by disk I/O latency caused by an over-burdened storage subsystem.