Which command helps view storage latency spikes by showing extended I/O statistics per device?

• A. iostat -xz 1
• B. iostat -x 1
• C. iotop
• D. dstat

Answer: (A)

To spot storage latency spikes you need a tool that shows per-device I/O statistics over time, including how long I/O requests wait. iostat with extended statistics provides per-device metrics like await (average wait time) and throughput. Adding the -z option hides devices with no activity, so you don’t waste time on idle disks, and using a 1-second interval makes spikes appear as they happen. Together, this lets you observe real-time latency behavior for each device, highlighting when latency spikes occur.

Other options aren’t as focused. iostat with extended statistics but without hiding idle devices includes inactive ones, adding noise. iotop shows I/O per process rather than per device latency details. dstat is broader and versatile but doesn’t emphasize per-device latency spikes as clearly as iostat with extended stats and a short update interval.