Which performance aspects should be reviewed to identify an I/O bottleneck in ONTAP?

• A. Latency, queue depth, IOPS, and throughput
• B. Temperature, fan speed, and chassis color
• C. Amount of free space only
• D. Network latency only

Answer: (A)

When identifying an I/O bottleneck in ONTAP, you examine latency, queue depth, IOPS, and throughput. Latency tells you how long each I/O operation takes; consistently high latency points to a bottleneck somewhere in the storage path, network, or compute that’s delaying responses. Queue depth shows how many I/Os are waiting to be processed; a chronically high queue depth means the system can’t keep up with the workload, signaling a processing or capacity constraint. IOPS measures the number of I/O operations the system handles per second; if IOPS are high but latency remains high, there’s contention or saturation affecting individual I/O completions. Throughput indicates the data transfer rate; hitting a throughput ceiling means the bandwidth between the storage and clients is maxed out, limiting performance regardless of IOPS or latency.

Other options don’t directly reflect I/O performance: environmental indicators like temperature and fans don’t diagnose throughput or latency; free space by itself doesn’t capture how fast data can be read or written; network latency alone misses how storage devices and controllers respond to requests. Together, the four metrics give a complete view of where the I/O path is slowing down.