VSP 360 · Array IOPS, past month16 Sep – 16 Oct 2026 · 5-minute samples · CEST

Array A's load doubled on 8 October and has stayed there. Array B has run the same 33,000 IOPS every day for a month.

Array A averaged about 7,100 IOPS a day in mid-September, rose through late September and stepped up to 15,000–17,000 from 8 October, with bursts above 30,000 every day. Array B never moved: daily averages between 32,600 and 36,500 IOPS, the same shape every hour. Array C, monitored since 25 September, idles at about 90 IOPS apart from short bursts, the largest 22,123 IOPS on 9 October.

2.1×

Array A growth

About 7,100 IOPS a day (16–20 Sep) to 15,000 (8–16 Oct)

33,500IOPS

Array B, every day

Daily averages 32,600–36,500 for 31 days

~90IOPS

Array C baseline

Bursts on 9, 10, 13–15 Oct

71,533IOPS

Highest 5 minutes

Array B, 18 Sep; twice its normal rate

Figure · One row per day, midnight to midnight

The step on Array A shows as the rows turning darker from 8 October; Array B is the same band, top to bottom

How to read a horizon chart: each row is one day's IOPS folded into three layers so it fits a thin strip. The lightest layer is the first band of IOPS, the middle shade the second, the darkest the third; anything above three bands is drawn in orange. Arrays A and B share one scale, 14,000 IOPS per band. Array C uses its own scale, 100 IOPS per band, because at the Array A and B scale it would not show; its rows are not comparable in height with the others. Numbers at the right are the day's average and its busiest 5 minutes. Blank stretches are missing data.