Skip to main content


I have a physical server with #raid 0 consist of two 2 tb hdd disks.
Recently I've started to receive warnings from my #prometheus about too high disk write latency (expr: (rate(node_disk_write_time_seconds_total[1m]) / rate(node_disk_writes_completed_total[1m]) > 0.1 and rate(node_disk_writes_completed_total[1m]) > 0) * on(instance) group_left (nodename) node_uname_info{nodename=~".+"}).

In alerts I see values around 0.200 to 0.550.
So my question is simple: is it really a warning and I should worry about my data, make backups, talk with DC staff about replacing my disk?
Should I check my disks for errors and slow sectors and if yes then how do I do this?
My server has Debian 12 and a bunch of different services on it, including a small internet radio, gitea, mail server and several docker containers with small stuff.

#Linux #server #askfedi