Snippets Groups Projects

Using prometheus to alert on

Dump the output of this into /var/lib/node_exporter/textfile_collector/mdadm-monitor.prom and then add this to your alerting rules:

# Mdadm monitor not running.  Last TestMessage event was 1800 seconds ago.
ALERT MdadmMonitorNotRunning
  IF mdadm_monitor{event="TestMessage"} < (time() - 1800)
  FOR 80m
  ANNOTATIONS {
    summary = "Mdadm Monitor has problems on {{$labels.instance}} failed.",
    description = "Mdadm Monitor has problems on {{$labels.instance}} failed.",
  }

# Mdadm monitor sees a problem.  An event has fired in last 1800 seconds.
ALERT MdadmMonitorErrorDetected
  IF mdadm_monitor{event!="TestMessage"} > (time() - 1800)
  FOR 30m
  ANNOTATIONS {
    summary = "Mdadm Monitor has seen an event on {{$labels.instance}}.",
    description = "Mdadm Monitor has an event ({{$labels.event}}) on {{$labels.instance}}.",
  }

mdadm-monitor.sh 274 B

#!/bin/bash

now=$(date '+s/$/ %s/')

cat << EOF
# HELP mdadm_monitor Follows events on the arrays.
# TYPE mdadm_monitor gauge
EOF
mdadm --monitor --scan --test --oneshot -p /bin/echo \
  | awk '{print("mdadm_monitor{event=\"" $1 "\",device=\"" $2 "\"}");}' \
  | sed "$now"

Please register or to comment