I use telegraf version 1.6.1 to gather vm processes metrics, when install by rpm, num_fds and io metrics are missing in the resulting output. i see telegraf github issues https://github.com/influxdata/telegraf/issues/517, https://github.com/influxdata/telegraf/issues/2403 , but that running telegraf as root solves the issue, another ansower to solve this issue?
I tried setting
CAP_SYS_ADMIN to telegraf using
setcap but it didn’t work. The only time it worked was when I ran it as root.