Icinga 2: Send SMS alerts only for CRITICAL status and its recoveries

Written by Claudio Kuenzler - 0 comments

Published on November 10th 2015 - Listed in Monitoring Nagios Icinga

---

Needed to find a way to configure Icinga 2, that only alerts with a CRITICAL state are sent by SMS (no WARNING or UNKNOWN states). However a recovery (OK) from a CRITICAL state should of course also be sent.

To handle this, I used the service.last_state and service.last_state_id macros, which contain exactly this information; what was the prior status of this service?

A special NotificationCommand definition for sending SMS alerts was added and it contains the last_state macros:

object NotificationCommand "notify-service-by-sms" {
        import "plugin-notification-command"
        command = [ SysconfDir + "/icinga2/scripts/sms-service-notification.sh" ]

        env = {
        NOTIFICATIONTYPE = "$notification.type$"
        SERVICEDESC = "$service.name$"
        HOSTALIAS = "$host.display_name$"
        HOSTADDRESS = "$address$"
        SERVICESTATE = "$service.state$"
        SERVICEPREVIOUSSTATE = "$service.last_state$"
        SERVICESTATEID = "$service.state_id$"
        SERVICEPREVIOUSSTATEID = "$service.last_state_id$"
        LONGDATETIME = "$icinga.long_date_time$"
        SHORTDATETIME = "$icinga.short_date_time$"
        SERVICEOUTPUT = "$service.output$"
        NOTIFICATIONAUTHORNAME = "$notification.author$"
        NOTIFICATIONCOMMENT = "$notification.comment$"
        HOSTDISPLAYNAME = "$host.display_name$"
        SERVICEDISPLAYNAME = "$service.display_name$"
        USEREMAIL = "$user.email$"
        USERPAGER = "$user.pager$"
        }
}

As you see, this NotificationCommand "notify-service-by-sms" will launch the script /etc/icinga2/scripts/sms-service-notification.sh and will send the enviornment variables mentioned to the script.

The script itself contains the logic when and how to send the SMS:

#!/bin/bash

# 20151109 Claudio: Only send SMS when $USERPAGER is set and when $SERVICESTATE is CRITICAL
# 20151110 Claudio: Also send OK (Recovery) when previous state was CRITICAL

sendsms() {
templatesms=`cat <<TEMPLATE
$NOTIFICATIONTYPE
Service: $SERVICEDESC on Host $HOSTALIAS is $SERVICESTATE
$SHORTDATETIME
TEMPLATE
`
/usr/bin/printf "%b" "$templatesms" | mailx $USERPAGER -- -f icinga@example.com
}

if ! [[ $USERPAGER = "" ]] ; then

  if [[ $SERVICESTATEID -eq 2 ]]; then sendsms
  elif [[ $SERVICESTATEID -eq 0 ]] && [[ $SERVICEPREVIOUSSTATEID -eq 2 ]]; then sendsms
  fi

fi

# Always send e-mail:

templatemail=`cat <<TEMPLATE
***** Icinga  *****

Notification Type: $NOTIFICATIONTYPE

Service: $SERVICEDESC
Host: $HOSTALIAS
Address: $HOSTADDRESS
State: $SERVICESTATE

Date/Time: $LONGDATETIME

Additional Info: $SERVICEOUTPUT

Comment: [$NOTIFICATIONAUTHORNAME] $NOTIFICATIONCOMMENT
TEMPLATE
`
/usr/bin/printf "%b" "$templatemail" | mailx -s "$NOTIFICATIONTYPE - $HOSTDISPLAYNAME - $SERVICEDISPLAYNAME is $SERVICESTATE" $USEREMAIL

Add a comment

Show form to leave a comment

Comments (newest first)

No comments yet.