Alerts


/etc/prometheus/alerts.yml > alerts
node_few_inodes_remaining (0 active)
alert: node_few_inodes_remaining
expr: 100 * (sum without(fstype, job, mountpoint) (node_filesystem_files_free{job="node-exporter",mountpoint="/"}) / sum without(fstype, job, mountpoint) (node_filesystem_files{job="node-exporter",mountpoint="/"})) < 5
for: 10m
labels:
  severity: S2 - major
annotations:
  description: |
    Seems that {{ $labels.instance }} is about to run out of free inodes (currently at {{ humanize $value }}%). If they run out, we'll start getting "no space left on device" errors and suffer data loss. Try SSH-ing into the server and running "docker system prune".
  summary: Only {{ humanize $value }}% of inodes remaining on {{ $labels.instance }}.
node_low_disk_space (0 active)
alert: node_low_disk_space
expr: 100 * (sum without(fstype, job, mountpoint) (node_filesystem_avail_bytes{device!~"rootfs",job="node-exporter",mountpoint="/"}) / sum without(fstype, job, mountpoint) (node_filesystem_size_bytes{device!~"rootfs",job="node-exporter",mountpoint="/"})) < 10
for: 10m
labels:
  severity: S2 - major
annotations:
  description: |
    Seems that {{ $labels.instance }} is about to run out of free space (currently at {{ humanize $value }}%). If it runs out, we might suffer data loss. Try SSH-ing into the server and running "docker system prune".
  summary: Only {{ humanize $value }}% of disk space remaining on {{ $labels.instance }}.