[Linux]/proc/diskstats description of what each column means and how disk usage is calculated

Linux system/proc directory is stored in the current system running state information, as a file to save, such as CPU information file Cpuinfo, memory information file meminfo, disk information file diskstats.

The meanings of each column in Diskstats are as follows:

1 0 SDA 505235 28461 7421219 6259625 2397331 3804539 49818500 25885599 0 6122393 32152888

Device number device read completed number of times read number of reads sector count read number of milliseconds write completed number of writes completed number of writes sector number of milliseconds the write operation took a millisecond number of input/output requests input/output operations spent in milliseconds input/output operation The weighted number of milliseconds to spend.

In addition to the non-cumulative value of the number of input/output requests being processed, other disk statistics are cumulative values.

The disk usage in Zabbix is collected from Diskstats and the disk usage is calculated as follows:

The difference in the number of milliseconds spent in input/output operations of two acquisitions/acquisition interval time

For example: the first acquisition input/output operation takes 90258834 milliseconds, 10 seconds after the acquisition of a value of 90258710

Then the disk usage is (90258710ms-90258834ms)/10*1000ms = 0.0124, which is 1.24%

[Linux]/proc/diskstats description of what each column means and how disk usage is calculated