Replacement for SMTP-check (supports SSL)

[frater]

I don't know what the "Simple" smtp-check does.
All I know is that it triggers more than it should and it doesn't support SSL.

I therefore wrote this small external script. Most of the work is done by "swaks". You can easily install it with yum or apt-get (apt-get install swaks).

Item
   Email (SMTP) server
   external
   smtpup[ {HOST.CONN1 ]
   interval 60s
   Show value "Service State"

mkdir /etc/zabbix/external
ln -s /usr/local/sbin/smtpup /etc/zabbix/external/smtpup

cat /usr/local/sbin/smtpup

#!/bin/bash
export PATH=${PATH}:/usr/local/sbin:/sbin:/usr/sbin:/bin:/usr/bin

# Script mainly used for Zabbix (http://www.zabbix.com) to check if an SMTP-server is up
# Zabbix itself has an smtp-check function, but due to its nature it gives many false positives.
# This script, mainly because it uses swaks, has a dialogue with the SMTP-server that's comparable
# with a real-life SMTP-session. It says Hello (HELO) and then the SMTP-server tells the things he
# understands. This script then gives a QUIT, because it just wants to test if the server is working#
# Put the script in /usr/local/sbin and make a symbolic link in /etc/zabbix/externalscripts
#
# ln -s /usr/local/sbin/smtpup /etc/zabbix/externalscripts/smptup
#
# The latter is important because it needs this to differentiate between executing it from
# command-line and executing it from zabbix (zabbix gives an extra parameter)

#
# An item in zabbix is like this:
#
#       Description:            Email (SMTPAUTH) server is responding
#       Type:                   external check
#       Key:                    smtpup[ {HOST.CONN1} 587 ]
#       Type of information:    Numeric (unsigned)
#       Data type:              Decimal
#       Update interval:        60
#       Show value:             Service state


# This Zabbix' workaround will lose the first paramater when executed from
# /etc/zabbix/externalscripts/
echo "${BASH_SOURCE}" | grep -q "zabbix" && shift 1

# Server
OPTIONS=
SERVER="$1"
PORT="$2"

# CONSTANTS
RETURN=0
QUITON=HELO
REPLY='^<. *250'
TIMEOUT=5

# Put your Internet FQDN-address here.
# If your IP also has its reverse DNS to that same FQDN it is even better.

# HELO=smtp.yourdomain.com
HELO=
[ -z "${HELO}" ] || OPTIONS="${OPTIONS} --helo ${HELO}"

if [ ! -z "${SERVER}" ] ; then
 [ -z ${PORT} ]         && PORT=25
 [ "${PORT}" = "465" ]  && OPTIONS="${OPTIONS} -tlsc"

 swaks ${OPTIONS} -q ${QUITON} -p ${PORT} --timeout ${TIMEOUT} -s ${SERVER} 2>/dev/null | grep -q "${REPLY}" && RETURN=1
fi
echo "$RETURN"

[frater]

Both having used this script for a while and monitoring an extreme slow server I decided to enhance the script so it makes full use of the 'TimeOut=' value in /etc/zabbix/zabbix_server.conf

Although I was using the "--timeout" value in swaks, it took significantly longer than the limit I imposed. This is not good with Zabbix, as it will turn off monitoring for this item when it is supposed to give an alarm bell.

I decided to write a workaround for it by running swaks in the background and have my own wait cycle.

More info can be found in the script itself.

I have added the template Standalone where I use this script and 2 other I wrote (httpspeed & certexpire)

#!/bin/bash
export PATH=${PATH}:/usr/local/sbin:/sbin:/usr/sbin:/bin:/usr/bin

# Script mainly used for Zabbix (http://www.zabbix.com) to check if an SMTP-server is up
# Zabbix itself has an smtp-check function, but due to its nature it gives many false positives.
# This script, mainly because it uses swaks, has a dialogue with the SMTP-server that's comparable
# with a real-life SMTP-session. It says Hello (HELO) and then the SMTP-server tells the things he
# understands. This script then gives a QUIT, because it just wants to test if the server is working#
# Put the script in /usr/local/sbin and make a symbolic link in /etc/zabbix/externalscripts
#
# ln -s /usr/local/sbin/smtpup /etc/zabbix/externalscripts/smptup
#
# The latter is important because it needs this to differentiate between executing it from
# command-line and executing it from zabbix (zabbix gives an extra parameter)

#
# An item in zabbix is like this:
#
#       Description:            Email (SMTPAUTH) server is responding
#       Type:                   external check
#       Key:                    smtpup[ {HOST.CONN1} 587 ]
#       Type of information:    Numeric (unsigned)
#       Data type:              Decimal
#       Update interval:        60
#       Show value:             Service state


# This Zabbix' workaround will lose the first paramater when executed from
# /etc/zabbix/externalscripts/

TIMEOUT=20
RETVAL=0

# If called by zabbix, handle some things different
if echo "${BASH_SOURCE}" | grep -q "zabbix" ; then
  # get rid of 1st parameter
  shift 1

  # Change TimeOut value to the one in /etc/zabbix/zabbix_server.conf
  ZABBIX_TIMEOUT=`grep -i '^Timeout' /etc/zabbix/zabbix_server.conf 2>/dev/null | awk -F= '{print $2}' | tr -cd '0-9'`
  if [ -z "${ZABBIX_TIMEOUT}" ] ; then
    TIMEOUT=3
  else
    # Let's take 1 second less than the one in /etc/zabbix/zabbix_server.conf and just hope to be in time
    TIMEOUT=$(( ${ZABBIX_TIMEOUT} - 1 ))
  fi
fi

# Server
OPTIONS=
SERVER="$1"
PORT="$2"

# CONSTANTS
QUITON=HELO
REPLY='^<. *250'

# Put your Internet FQDN-address here.
# If your IP also has its reverse DNS to that same FQDN it is even better.

# HELO=smtp.yourdomain.com
HELO=

[ -z "${HELO}" ] || OPTIONS="${OPTIONS} --helo ${HELO}"

if [ ! -z "${SERVER}" ] ; then
 [ -z ${PORT} ]         && PORT=25
 [ "${PORT}" = "465" ]  && OPTIONS="${OPTIONS} -tlsc"

 SCRATCH=`mktemp`

 # Somehow swaks doesn't properly obey TimeOuts..  I'm running it in background and use my own timeloop

 # suppress terminal output
 exec 2>/dev/null
 # Start swaks in background
 swaks --timeout ${TIMEOUT} ${OPTIONS} -q ${QUITON} -p ${PORT} -s ${SERVER} 2>/dev/null >${SCRATCH} &

 # little offset.... If it responds fast, it doesn't need to enter the loop
 sleep .2

 # Wait for swaks
 n=1
 while ! grep -q "${REPLY}" ${SCRATCH} ; do
  sleep .49
  grep -q "${REPLY}" ${SCRATCH} && break
  sleep .5
  [ $n -ge ${TIMEOUT} ] && break
  let n++
 done

 if grep -q "${REPLY}" ${SCRATCH} ; then
   RETVAL=1
 fi

 # If it didn't sign off, we should better kill it
 if ! grep -q "QUIT" ${SCRATCH} ; then
  # Too late you lazy bastard, I might as well kill you...
  kill %1 2>/dev/null
 fi

 rm -f ${SCRATCH} 2>/dev/null
fi
echo "$RETVAL"