Thursday, April 7, 2016

Simple monitoring

I was searhing for a simple monitoring tool but most tools out there are very complex. We use 3rd party virtual private servers now so hardware monitoring is not our business. We're interested in monitoring two things:
  1. Free disk space
  2. CPU load
So I use two simple scripts that send out an e-mail alert when the diskspace is low or the CPU load it too high.

You can create the two scripts below and edit the MAILTO and TRIGGER variables to suit your needs /usr/local/bin/


host=`hostname -f`
load=`cat /proc/loadavg | awk '{print $2}'`
response=`echo | awk -v T=$TRIGGER -v L=$load 'BEGIN{if ( L > T){ print "greater"}}'`
if [[ $response = "greater" ]]
        #capture top command output for the mail body
        body=`top -n 1 -b`

        echo "$body\n."|mail -s "High load on $host - [ $load ]" $MAILTO


HOST=`hostname -f`
CURRENT=$(df / | grep / | awk '{ print $5}' | sed 's/%//g')

if [ "$CURRENT" -gt "$TRIGGER" ] ; then

    mail -s "Disk Space Alert for $HOST" $MAILTO << EOF
The root partition of '$HOST' remaining free space is critically low. Used: $CURRENT%

Make these two scripts executable:
chmod +x /usr/local/bin/ 
chmod +x /usr/local/bin/
Now schedule these script in system cron jobs:
# Check server load and diskspace
* * * * * root /usr/local/bin/
0 7,16 * * * root /usr/local/bin/
This will check the server load every minute and disk space at 7am and 4pm. Finally it's wise to test if mails sent from console arrive. Change the e-mail address in this command and run it:
echo "."|mail -s "Test mail from console"