When using conky to monitor disk I/O, the default scale is dynamic. This makes it unclear how much I/O is actually happening – a high bar will appear even when little activity is occurring due to the scaling.

Unfortunately the scale feature is not documented well.

Using a non-logarithmic scale of 1000 seems to present reasonable results:

$(diskiograph_read /dev/sda 00ff00 ff0000 1000 -t}