Matejunkie » gplv2

PNP template for check_apache2.sh

Mike Adolphs — Tue, 08 Sep 2009 20:07:52 +0000

This is a PNP template which works quite well for me when there’s need to identify possible bottlenecks or unexpected behavior of the Apache webserver. It shows the CPU usage, req/sec, byte/req and the amount of busy/idle workers in a single graph. Feel free to copy the template from below or head over to Nagios Exchange.

#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

#   PNP Template for check_apache2.sh
#   Author: Mike Adolphs (http://www.matejunkie.com/)

$opt[1] = "--vertical-label \"Various stats \" -l 0 -r --title \"Apache metrics for $hostname / $servicedesc\" ";
$def[1]  = "DEF:cpu=$rrdfile:$DS[1]:AVERAGE " ;
$def[1] .= "DEF:reqpsec=$rrdfile:$DS[2]:AVERAGE " ;
$def[1] .= "DEF:bytepreq=$rrdfile:$DS[4]:AVERAGE " ;
$def[1] .= "DEF:wbusy=$rrdfile:$DS[5]:AVERAGE " ;
$def[1] .= "DEF:widle=$rrdfile:$DS[6]:AVERAGE " ;
 
$opt[2] .= "--vertical-label \"Byte/sec \" -l 0 -r --title \"Apache metrics for $hostname / $servicedesc\" ";
$def[2] .= "DEF:bytepsec=$rrdfile:$DS[3]:AVERAGE " ;
 
$def[1] .= "COMMENT:\"\\t\\t\\t\\tLAST\\t\\tAVERAGE\\t\\tMAX\\n\" " ;
 
$def[1] .= "LINE2:cpu#008000:\"CPU usage [%]\\t \" " ;
$def[1] .= "GPRINT:cpu:LAST:\"%6.0lf\\t\" " ;
$def[1] .= "GPRINT:cpu:AVERAGE:\" %6.0lf\\t\" " ;
$def[1] .= "GPRINT:cpu:MAX:\" %6.0lf\\n\" " ;
 
$def[1] .= "LINE2:reqpsec#0C64E8:\"Requests/sec\\t \" " ;
$def[1] .= "GPRINT:reqpsec:LAST:\"%6.0lf\\t\" " ;
$def[1] .= "GPRINT:reqpsec:AVERAGE:\" %6.0lf\\t\" " ;
$def[1] .= "GPRINT:reqpsec:MAX:\" %6.0lf\\n\" " ;
 
$def[1] .= "LINE2:bytepreq#FFA500:\"Byte/req\\t\\t \" " ;
$def[1] .= "GPRINT:bytepreq:LAST:\"%6.0lf\\t\" " ;
$def[1] .= "GPRINT:bytepreq:AVERAGE:\" %6.0lf\\t\" " ;
$def[1] .= "GPRINT:bytepreq:MAX:\" %6.0lf\\n\" " ;
 
$def[1] .= "LINE2:wbusy#1CC8E8:\"Busy workers\\t \" " ;
$def[1] .= "GPRINT:wbusy:LAST:\"%6.0lf\\t\" " ;
$def[1] .= "GPRINT:wbusy:AVERAGE:\" %6.0lf\\t\" " ;
$def[1] .= "GPRINT:wbusy:MAX:\" %6.0lf\\n\" " ;
 
$def[1] .= "LINE2:widle#E80C8C:\"Idle workers\\t \" " ;
$def[1] .= "GPRINT:widle:LAST:\"%6.0lf\\t\" " ;
$def[1] .= "GPRINT:widle:AVERAGE:\" %6.0lf\\t\" " ;
$def[1] .= "GPRINT:widle:MAX:\" %6.0lf\\n\" " ;
 
$def[2] .= "COMMENT:\"\\t\\t\\t\\tLAST\\t\\tAVERAGE\\t\\tMAX\\n\" " ;
 
$def[2] .= "LINE2:bytepsec#E80C3E:\"Byte/sec\\t\\t \" " ;
$def[2] .= "GPRINT:bytepsec:LAST:\"%6.0lf\\t\" " ;
$def[2] .= "GPRINT:bytepsec:AVERAGE:\" %6.0lf\\t\" " ;
$def[2] .= "GPRINT:bytepsec:MAX:\" %6.0lf\\n\" " ;

Update 1.3: Apache check plugin for Nagios

Mike Adolphs — Mon, 07 Sep 2009 22:49:16 +0000

Long time no see. Sorry for the long downtime, but I’ve been on vacation for quite some time and spent my time out of reach of any electronical device except my phone. Anyway, strength and motivation for work during spare time is back again! Expect a few updates of various plugins I’ve written in the past within the next weeks.

For the beginning here’s the revised version of the Apache check plugin for Nagios. It now doesn’t write temporary files anymore, the code is cleaner and it should be a bit faster. Note that I haven’t done much testing for now (will be done tomorrow and on Wednesday). Therefore it might won’t work out of the box for you. At least in my environments it works well. Let me know if you have any problems or if you stumble upon something which is completely nonsense.

Copy’n'paste below or svn co via svn.matejunkie.com/nagios-plugins/trunk/check_apache/.

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.3,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "Description:"
    echo "$PROGNAME is a Nagios plugin to check the Apache's server status."
    echo "It monitors requests per second, bytes per second/request, "
    echo "amount of busy/idle workers and its CPU load."
    echo ""
    echo "Example call:"
    echo "./$PROGNAME -H localhost -P 80 -t 3 -b /usr/sbin -p /var/run \\"
    echo "-n apache2.pid -s status_page [-S] [-R] [-wr] 100 [-cr] 250"
    echo ""
    echo "Options:"
    echo "  -H|--hostname)"
    echo "    Sets the hostname. Default is: localhost"
    echo "  -P|--port)"
    echo "    Sets the port. Default is: 80"
    echo "  -t|--timeout)"
    echo "    Sets a timeout within the server's status page must've been"
    echo "    accessed. Otherwise the check will go into an error state."
    echo "    Default is: 3"
    echo "  -b|--binary-path)"
    echo "    Sets the path to the apache binary. Used for getting Apache's"
    echo "    CPU load. Default is: /usr/sbin"
    echo "  -p|--pid-path)"
    echo "    Path to Apache's pid file. Default is: /var/run"
    echo "  -n|--pid-name)"
    echo "    Name of Apache's pid file. Default is: apache2.pid"
    echo "  -s|--status-page)"
    echo "    Defines the name of the status page. Default is: server-status"
    echo "  -R|--remote-server)"
    echo "    Disabled the pid check so that remote Apaches can be queried."
    echo "    Default is: off"
    echo "  -S|--secure)"
    echo "    Enables HTTPS (no certificate check though). Default is: off"
    echo "  -wr|--warning-req)"
    echo "    Sets a warning level for requests per second. Default is: off"
    echo "  -cr|--critical-req)"
    echo "    Sets a critical level for requests per second. Default is: off"
    exit $ST_UK
}
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
hostname="localhost"
port=80
remote_srv=0
path_binary="/usr/sbin"
path_pid="/var/run"
name_pid="apache2.pid"
status_page="server-status"
timeout=3
secure=0
running=0
 
wcdiff_req=0
wclvls_req=0
 
while test -n "$1"; do
    case "$1" in
        --help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --hostname|-H)
            hostname=$2
            shift
            ;;
        --port|-P)
            port=$2
            shift
            ;;
        --timeout|-t)
            timeout=$2
            shift
            ;;
        --remote-server|-R)
            remote_srv=1
            ;;
        --binary_path|-b)
            path_binary=$2
            shift
            ;;
        --pid_path|-p)
            path_pid=$2
            shift
            ;;
        --pid_name|-n)
            name_pid=$2
            shift
            ;;
        --status-page|-s)
            status_page=$2
            shift
            ;;
        --secure|-S)
            secure=1
            ;;
        --warning-req|-wr)
            warn_req=$2
            shift
            ;;
        --critical-req|-cr)
            crit_req=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
    esac
    shift
done
 
# check functions
val_wcdiff_req() {
    if [ ! -z "$warn_req" -a ! -z "$crit_req" ]
    then
        wclvls_req=1
 
        if [ ${warn_req} -gt ${crit_req} ]
        then
            wcdiff_req=1
        fi
    elif [ ! -z "$warn_req" -a -z "$crit_req" ]
    then
        wcdiff_req=2
    elif [ -z "$warn_req" -a ! -z "$crit_req" ]
    then
        wcdiff_req=3
    fi
}
 
check_pid() {
    if [ -f "$path_pid/$name_pid" ]
    then
        retval=0
    else
        retval=1
    fi
}
 
check_processes() {
    if [ $1 -lt 1 ]
    then
        echo "UNKNOWN - Your Apache server seems not to run. Is your Nagios \
privileged to run 'ps ax' and is the Apache2 binary really located in \
$path_binary?"
        exit $ST_UK
    fi
}
 
check_output() {
    stat_output=`stat -c %s ${output_dir}/server-status`
    if [ "$stat_output" = 0 ]
    then
        echo "UNKNOWN - Local copy of server-status is empty. Are we \
allowed to access http://${hostname}:${port}/server-status?"
        exit $ST_UK
    fi
}
 
# get functions
get_status() {
    if [ "$secure" = 1 ]
    then
        server_status1=`wget -qO- --no-check-certificate -t 3 \
-T ${timeout} https://${hostname}:${port}/${status_page}?auto`
    sleep 1
        server_status2=`wget -qO- --no-check-certificate -t 3 \
-T ${timeout} https://${hostname}:${port}/${status_page}?auto`
    else
        server_status1=`wget -qO- -t 3 -T ${timeout} \
http://${hostname}:${port}/${status_page}?auto`
        sleep 1
        server_status2=`wget -qO- -t 3 -T ${timeout} \
http://${hostname}:${port}/${status_page}?auto`
    fi
}
 
get_vals() {
    cpu_load="$(cpu_load=0; ps -Ao pcpu,args | grep "$path_binary/apache2" \
| awk '{print $1}' | while read line
    do
        cpu_load=`echo "scale=3; $cpu_load + $line" | bc -l`
    echo $cpu_load
    done)"
    cpu_load=`echo $cpu_load | awk '{print $NF}' | sed 's/^\./0./'`
 
    tmp1_req_psec=`echo ${server_status1} | awk '{print $3}'`
    tmp2_req_psec=`echo ${server_status2} | awk '{print $3}'`
    req_psec=`echo "scale=2; ${tmp2_req_psec} - ${tmp1_req_psec}" | bc -l \
| sed 's/^\./0./'`
 
    bytes_psec=`echo ${server_status1} | awk '{print $14}' | sed 's/^\./0./'`
    bytes_preq=`echo ${server_status1} | awk '{print $16}' | sed 's/^\./0./'`
    wkrs_busy=`echo ${server_status1} | awk '{print $18}' | sed 's/^\./0./'`
    wkrs_idle=`echo ${server_status1} | awk '{print $20}' | sed 's/^\./0./'`
}
 
do_output() {
    output="Apache serves $req_psec Requests per second with an average CPU \
utilization of $cpu_load%. Busy workers: $wkrs_busy, idle: $wkrs_idle"
}
 
do_perfdata() {
    perfdata="'cpu_load'=$cpu_load 'req_psec'=$req_psec \
'bytes_psec'=$bytes_psec 'bytes_preq'=$bytes_preq 'workers_busy'=$wkrs_busy \
'workers_idle'=$wkrs_idle"
}
 
# Let's go
val_wcdiff_req
 
if [ "$wcdiff_req" = 1 ]
then
    echo "Please adjust your warning/critical thresholds. The warning must \
be lower than the critical level!"
    exit $ST_UK
elif [ "$wcdiff_req" = 2 ]
then
    echo "Please also set a critical value when you want to use \
warning/critical thresholds!"
    exit $ST_UK
elif [ "$wcdiff_req" = 3 ]
then
    echo "Please also set a warning value when you want to use \
warning/critical thresholds!"
    exit $ST_UK
else
    if [ "$remote_srv" = 0 ]
    then
        running=`check_pid`
        check_pid $running
    fi
 
    get_status
    get_vals
 
    do_output
    do_perfdata
 
    if [ ${wclvls_req} = 1 ]
    then
        if [ ${req_psec} -ge ${warn_req} -a ${req_psec} -lt ${crit_req} ]
        then
            echo "WARNING - ${output} | ${perfdata}"
            exit $ST_WR
        elif [ ${req_psec} -ge ${crit_req} ]
        then
            echo "CRITICAL - ${output} | ${perfdata}"
        exit $ST_CR
        else
            echo "OK - ${output} | ${perfdata}"
            exit $ST_OK
        fi
    else
        echo "OK - ${output} | ${perfdata}"
        exit $ST_OK
    fi
fi

Update 1.1: nginx Check Plugin for Nagios

Mike Adolphs — Tue, 16 Jun 2009 07:51:24 +0000

Last week I’ve got an E-Mail by Jordi, who’s blogging at terraquis.net, with a few suggestions to the nginx check plugin I’ve recently released. He made two significant changes to the plugin:

His version runs with bash only, but except bc all external calls have been replaced by bash built-ins
There’s no temporary file anymore which has to be written to a temporary directory, nginx_status page output is saved into a variable

The plugin therefore runs much faster than my initial version of the plugin: 14-16ms plus the 1 second sleep interval between the two nginx_status checks.

And then I realized that writing output to a temporary file is just one thing: plain stupid. Why writing files when an information can be stored in a variable where it stays in memory?
Well, I can’t really say why I walked this way when I’ve written the first plugins, but I eventually had the right idea when I’ve written the Hadoop DFS check. There I already store the information within variables. The only guess why I did this is that I’ve assumed that storing longer output in variables just would be too much.

But anyway, there has to be room for improvements, right? Therefore, expect a few reworks within the next days, …

The script

Below is the revised script I’ve already uploaded yesterday to Nagios and MonitoringExchange. It runs much faster than the initial version and doesn’t write files anymore. You may also stick to the output below or do the SVN way.

user@host: ~ $ svn co svn://svn.matejunkie.com/nagios-plugins/stable/check_nginx/ check_nginx/

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.1,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
hostname="localhost"
port=80
path_pid=/var/run
name_pid="nginx.pid"
status_page="nginx_status"
pid_check=1
secure=0
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to check whether nginx is running."
    echo "It also parses the nginx's status page to get requests and"
    echo "connections per second as well as requests per connection. You"
    echo "may have to alter your nginx configuration so that the plugin"
    echo "can access the server's status page."
    echo "The plugin is highly configurable for this reason. See below for"
    echo "available options."
    echo ""
    echo "$PROGNAME -H localhost -P 80 -p /var/run -n nginx.pid "
	echo "  -s nginx_statut -o /tmp [-w INT] [-c INT] [-S] [-N]"
    echo ""
    echo "Options:"
    echo "  -H/--hostname)"
    echo "     Defines the hostname. Default is: localhost"
    echo "  -P/--port)"
    echo "     Defines the port. Default is: 80"
    echo "  -p/--path-pid)"
    echo "     Path where nginx's pid file is being stored. You might need"
    echo "     to alter this path according to your distribution. Default"
    echo "     is: /var/run"
    echo "  -n/--name_pid)"
    echo "     Name of the pid file. Default is: nginx.pid"
    echo "  -N/--no-pid-check)"
    echo "     Turn this on, if you don't want to check for a pid file"
    echo "     whether nginx is running, e.g. when you're checking a"
    echo "     remote server. Default is: off"
    echo "  -s/--status-page)"
    echo "     Name of the server's status page defined in the location"
    echo "     directive of your nginx configuration. Default is:"
    echo "     nginx_status"
    echo "  -S/--secure)"
    echo "     In case your server is only reachable via SSL, use this"
    echo "     this switch to use HTTPS instead of HTTP. Default is: off"
    echo "  -w/--warning)"
    echo "     Sets a warning level for requests per second. Default is: off"
    echo "  -c/--critical)"
    echo "     Sets a critical level for requests per second. Default is:"
	echo "     off"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        -help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --hostname|-H)
            hostname=$2
            shift
            ;;
        --port|-P)
            port=$2
            shift
            ;;
        --path-pid|-p)
            path_pid=$2
            shift
            ;;
        --name-pid|-n)
            name_pid=$2
            shift
            ;;
        --no-pid-check|-N)
            pid_check=0
            ;;
        --status-page|-s)
            status_page=$2
            shift
            ;;
        --secure|-S)
            secure=1
            ;;
        --warning|-w)
            warning=$2
            shift
            ;;
        --critical|-c)
            critical=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
        esac
    shift
done
 
get_wcdiff() {
    if [ ! -z "$warning" -a ! -z "$critical" ]
    then
        wclvls=1
 
        if [ ${warning} -ge ${critical} ]
        then
            wcdiff=1
        fi
    elif [ ! -z "$warning" -a -z "$critical" ]
    then
        wcdiff=2
    elif [ -z "$warning" -a ! -z "$critical" ]
    then
        wcdiff=3
    fi
}
 
val_wcdiff() {
    if [ "$wcdiff" = 1 ]
    then
        echo "Please adjust your warning/critical thresholds. The warning \
must be lower than the critical level!"
        exit $ST_UK
    elif [ "$wcdiff" = 2 ]
    then
        echo "Please also set a critical value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    elif [ "$wcdiff" = 3 ]
    then
        echo "Please also set a warning value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    fi
}
 
check_pid() {
    if [ -f "$path_pid/$name_pid" ]
    then
        retval=0
    else
        retval=1
    fi
}
 
get_status() {
    if [ "$secure" = 1 ]
    then
        wget_opts="-O- -q -t 3 -T 3 --no-check-certificate"
        out1=`wget ${wget_opts} http://${hostname}:${port}/${status_page}`
        sleep 1
        out2=`wget ${wget_opts} http://${hostname}:${port}/${status_page}`
    else        
        wget_opts="-O- -q -t 3 -T 3"
        out1=`wget ${wget_opts} http://${hostname}:${port}/${status_page}`
        sleep 1
        out2=`wget ${wget_opts} http://${hostname}:${port}/${status_page}`
    fi
 
    if [ -z "$out1" -o -z "$out2" ]
    then
        echo "UNKNOWN - Local copy/copies of $status_page is empty."
        exit $ST_UK
    fi
}
 
get_vals() {
    tmp1_reqpsec=`echo ${out1}|awk '{print $10}'`
    tmp2_reqpsec=`echo ${out2}|awk '{print $10}'`
    reqpsec=`expr $tmp2_reqpsec - $tmp1_reqpsec`
 
    tmp1_conpsec=`echo ${out1}|awk '{print $9}'`
    tmp2_conpsec=`echo ${out2}|awk '{print $9}'`
    conpsec=`expr $tmp2_conpsec - $tmp1_conpsec`
 
    reqpcon=`echo "scale=2; $reqpsec / $conpsec" | bc -l`
    if [ "$reqpcon" = ".99" ]
    then
        reqpcon="1.00"
    fi
}
 
do_output() {
    output="nginx is running. $reqpsec requests per second, $conpsec \
connections per second ($reqpcon requests per connection)"
}
 
do_perfdata() {
    perfdata="'reqpsec'=$reqpsec 'conpsec'=$conpsec 'conpreq'=$reqpcon"
}
 
# Here we go!
get_wcdiff
val_wcdiff
 
if [ ${pid_check} = 1 ]
then
    check_pid
    if [ "$retval" = 1 ]
    then
        echo "There's no pid file for nginx. Is nginx running? Please \
also make sure whether your pid path and name is correct."
        exit $ST_CR
    fi
fi
 
get_status
get_vals
do_output
do_perfdata
 
if [ -n "$warning" -a -n "$critical" ]
then
    if [ "$reqpsec" -ge "$warning" -a "$reqpsec" -lt "$critical" ]
    then
        echo "WARNING - ${output} | ${perfdata}"
	exit $ST_WR
    elif [ "$reqpsec" -ge "$critical" ]
    then
        echo "CRITICAL - ${output} | ${perfdata}"
	exit $ST_CR
    else
        echo "OK - ${output} | ${perfdata} ]"
	exit $ST_OK
    fi
else
    echo "OK - ${output} | ${perfdata}"
    exit $ST_OK
fi

Jordi’s script

This is Jordi’s bash built-in (plus bc) version which runs even faster. Feel free to copy the output below.

#!/bin/bash
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
hostname="localhost"
port=80
path_pid=/var/run
name_pid="nginx.pid"
status_page="nginx_status"
pid_check=1
secure=0
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to check whether nginx is running."
    echo "It also parses the nginx's status page to get requests and"
    echo "connections per second as well as requests per connection. You"
    echo "may have to alter your nginx configuration so that the plugin"
    echo "can access the server's status page."
    echo "The plugin is highly configurable for this reason. See below for"
    echo "available options."
    echo ""
    echo "$PROGNAME -H localhost -P 80 -p /var/run -n nginx.pid "
	echo "  -s nginx_statut -o /tmp [-w INT] [-c INT] [-S] [-N]"
    echo ""
    echo "Options:"
    echo "  -H/--hostname)"
    echo "     Defines the hostname. Default is: localhost"
    echo "  -P/--port)"
    echo "     Defines the port. Default is: 80"
    echo "  -p/--path-pid)"
    echo "     Path where nginx's pid file is being stored. You might need"
    echo "     to alter this path according to your distribution. Default"
    echo "     is: /var/run"
    echo "  -n/--name_pid)"
    echo "     Name of the pid file. Default is: nginx.pid"
    echo "  -N/--no-pid-check)"
    echo "     Turn this on, if you don't want to check for a pid file"
    echo "     whether nginx is running, e.g. when you're checking a"
    echo "     remote server. Default is: off"
    echo "  -s/--status-page)"
    echo "     Name of the server's status page defined in the location"
    echo "     directive of your nginx configuration. Default is:"
    echo "     nginx_status"
    echo "  -o/--output-directory)"
    echo "     Unneeded. The option is preserved for backwards compatibility."
    echo "  -S/--secure)"
    echo "     In case your server is only reachable via SSL, use this"
    echo "     this switch to use HTTPS instead of HTTP. Default is: off"
    echo "  -w/--warning)"
    echo "     Sets a warning level for requests per second. Default is: off"
    echo "  -c/--critical)"
    echo "     Sets a critical level for requests per second. Default is:"
	echo "     off"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        -help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --hostname|-H)
            hostname=$2
            shift
            ;;
        --port|-P)
            port=$2
            shift
            ;;
        --path-pid|-p)
            path_pid=$2
            shift
            ;;
        --name-pid|-n)
            name_pid=$2
            shift
            ;;
        --no-pid-check|-N)
            pid_check=0
            ;;
        --status-page|-s)
            status_page=$2
            shift
            ;;
        --output-directory|-o)
            output_dir=$2
            shift
            ;;
        --secure|-S)
            secure=1
            ;;
        --warning|-w)
            warning=$2
            shift
            ;;
        --critical|-c)
            critical=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
        esac
    shift
done
 
get_wcdiff() {
    if [ ! -z "$warning" -a ! -z "$critical" ]
    then
        wclvls=1
 
        if [ ${warning} -gt ${critical} ]
        then
            wcdiff=1
        fi
    elif [ ! -z "$warning" -a -z "$critical" ]
    then
        wcdiff=2
    elif [ -z "$warning" -a ! -z "$critical" ]
    then
        wcdiff=3
    fi
}
 
val_wcdiff() {
    if [ "$wcdiff" = 1 ]
    then
        echo "Please adjust your warning/critical thresholds. The warning \
must be lower than the critical level!"
        exit $ST_UK
    elif [ "$wcdiff" = 2 ]
    then
        echo "Please also set a critical value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    elif [ "$wcdiff" = 3 ]
    then
        echo "Please also set a warning value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    fi
}
 
check_pid() {
    if [ -f "$path_pid/$name_pid" ]
    then
        retval=0
    else
        retval=1
    fi
}
 
get_status() {
    url="http://${hostname}:${port}/${status_page}"
 
    if [ "$secure" = 1 ] ; then
        wget_opts="-O- -q -t 3 -t 3 --no-check-certificate"
    else
        wget_opts="-O- -q -t 3 -t 3"
    fi
 
    out1=$(wget ${wget_opts} ${url})
    sleep 1
    out2=$(wget ${wget_opts} ${url})
 
    # Both commands must return some content.
    if [ -z "${out1}" -o -z "${out2}" ] ; then
        echo "UNKNOWN - Local copy/copies of $status_page is empty."
        exit $ST_UK
    fi
}
 
get_vals() {
    # Convert into an array.
    values1=(${out1})
    values2=(${out2})
 
    reqpsec=$(( ${values2[9]} - ${values1[9]} ))
    conpsec=$(( ${values2[8]} - ${values1[8]} ))
 
    reqpcon=`echo "scale=2; $reqpsec / $conpsec" | bc -l`
    if [ "$reqpcon" = ".99" ]
    then
        reqpcon="1.00"
    fi
}
 
do_output() {
    output="nginx is running. $reqpsec requests per second, $conpsec \
connections per second ($reqpcon requests per connection)"
}
 
do_perfdata() {
    perfdata="'reqpsec'=$reqpsec 'conpsec'=$conpsec 'conpreq'=$reqpcon"
}
 
# Here we go!
get_wcdiff
val_wcdiff
 
if [ ${pid_check} = 1 ]
then
    check_pid
    if [ "$retval" = 1 ]
    then
        echo "There's no pid file for nginx. Is nginx running? Please \
also make sure whether your pid path and name is correct."
        exit $ST_CR
    fi
fi
 
get_status
get_vals
do_output
do_perfdata
 
if [ -n "$warning" -a -n "$critical" ]
then
    if [ "$reqpsec" -ge "$warning" -a "$reqpsec" -lt "$critical" ]
    then
        echo "WARNING - ${output} | ${perfdata}"
	exit $ST_WR
    elif [ "$reqpsec" -ge "$critical" ]
    then
        echo "CRITICAL - ${output} | ${perfdata}"
	exit $ST_CR
    else
        echo "OK - ${output} | ${perfdata} ]"
	exit $ST_OK
    fi
else
    echo "OK - ${output} | ${perfdata}"
    exit $ST_OK
fi

License

Nagios Exchange opens up

Mike Adolphs — Tue, 09 Jun 2009 11:45:46 +0000

Just a short note from my side:
A week ago Ethan launched Nagios Exchange, an addon and plugin repository similar to MonitoringExchange, which is part of the whole Nagios website framework. I began to add all of my plugins to it as well, but due to the approval process it might take a while till they’re downloadable. I’ll let you know when it’s done.

Nevertheless, from now on you’re even more free from where to download the plugins I’ve written. From now on they’re available via this blog, via SVN, via MonitoringExchange and via Nagios Exchange. I’ll update them on all sources.

Hadoop DFS Check Plugin for Nagios

Mike Adolphs — Tue, 02 Jun 2009 22:46:45 +0000

Since I’m currently occupied with kicking off a several terabytes large Hadoop cluster I thought it’d be a good idea to provide a few Nagios plugins related to Hadoop.

Just in case you don’t know: Hadoop is a free Java software framework that supports data intensive distributed applications. It got its own filesystem called HDFS which scales up to petabytes of storage running on top of the operating system’s filesystem. Hadoop is inspired by Google’s MapReduce and GFS, the Google File System.

Yahoo for example uses several Hadoop Clusters with 100.000 CPUs total in 20.000 boxes (2.000 boxes per cluster) according to the official Hadoop website. If you haven’t checked it out yet, do so! This is – sorry for that – crazy shit and it’s working!
For the beginning I’d recommend reading the Wikipedia article, the offical documentation and, of course, watching the Hadoop related lectures over at Cloudera.

Anyway, the first plugin is almost finished and checks the amount of running DataNodes and the capacity of the DFS. It’s working, but I want to add a few more features so that the plugin become more flexible. Nevertheless, feel free to give it a try if you’re in the mood.

The Script

Hadoop is completely operated by the hadoop user and since it got its own filesystem it got its own permissions as well. To keep things clean and safe we won’t give hadoop related permissions to the Nagios user. We’ll rather enable the Nagios user to run a small shell script as root which contains a command ran by the hadoop user to get information about the cluster.

Add the following small script to a directory of your choice, e.g. /usr/local/sbin:

#!/bin/sh
su -s /bin/bash - hadoop -c 'hadoop dfsadmin -report'

Afterwards alter its permissions so that it’s only read-, write and accessible by root:

user@host: ~ $ sudo chmod 700 /usr/local/sbin/get-dfsreport.sh

Then enable the Nagios user to sudo run the script via /etc/sudoers (or better visudo):

nagios          ALL=(ALL)       NOPASSWD: /usr/local/sbin/get-dfsreport.sh

This is it for the prerequisites. You may then run the script provided for copy’n'pasting below (or svn co, your choice).
If you’ve chosen another directory than /usr/local/sbin you have to provide the path via -s/–path-sh when running the script.

user@host: ~ $ svn co svn://svn.matejunkie.com/nagios-plugins/stable/check_hadoop-dfs/ check_hadoop-dfs/

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
path_sh="/usr/local/sbin"
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to check the status of HDFS, Hadoop's"
    echo "underlying, redundant, distributed file system."
    echo ""
    echo "$PROGNAME -s /usr/local/sbin [-w 10] [-c 5]"
    echo ""
    echo "Options:"
    echo "  -s|--path-sh)"
    echo "     Path to the shell script that is mentioned in the"
    echo "     documentation. Default is: /usr/local/sbin"
    echo "  -w|--warning)"
    echo "     Defines the warning level for available datanodes. Default"
    echo "     is: off"
    echo "  -c|--critical)"
    echo "     Defines the critical level for available datanodes. Default"
    echo "     is: off"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        -help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --path-sh|-s)
            path_sh=$2
            shift
            ;;
        --warning|-w)
            warning=$2
            shift
            ;;
        --critical|-c)
            critical=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
        esac
    shift
done
 
get_wcdiff() {
    if [ ! -z "$warning" -a ! -z "$critical" ]
    then
        wclvls=1
 
        if [ ${warning} -lt ${critical} ]
        then
            wcdiff=1
        fi
    elif [ ! -z "$warning" -a -z "$critical" ]
    then
        wcdiff=2
    elif [ -z "$warning" -a ! -z "$critical" ]
    then
        wcdiff=3
    fi
}
 
val_wcdiff() {
    if [ "$wcdiff" = 1 ]
    then
        echo "Please adjust your warning/critical thresholds. The warning \
must be higher than the critical level!"
        exit $ST_UK
    elif [ "$wcdiff" = 2 ]
    then
        echo "Please also set a critical value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    elif [ "$wcdiff" = 3 ]
    then
        echo "Please also set a warning value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    fi
}
 
get_vals() {
    tmp_vals=`sudo ${path_sh}/get-dfsreport.sh`
    dn_avail=`echo -e "$tmp_vals" | grep -m1 "Datanodes available:" | awk '{print $3}'`
    dfs_used=`echo -e "$tmp_vals" | grep -m1 "Used raw bytes:" | awk '{print $4}'`
    dfs_used=`expr ${dfs_used} / 1024 / 1024`
    dfs_used_p=`echo -e "$tmp_vals" | grep -m1 "% used:" | awk '{print $3}'`
    dfs_total=`echo -e "$tmp_vals" | grep -m1 "Total raw bytes:" | awk '{print $4}'`
    dfs_total=`expr ${dfs_total} / 1024 / 1024`
}
 
do_output() {
    output="Datanodes up and running: ${dn_avail}, DFS total: \
${dfs_total} MB, DFS used: ${dfs_used} MB (${dfs_used_p})"
}
 
do_perfdata() {
    perfdata="'datanodes_available'=${dn_avail} 'dfs_total'=${dfs_total} \
'dfs_used'=${dfs_used}"
}
 
# Here we go!
get_wcdiff
val_wcdiff
 
get_vals
 
do_output
do_perfdata
 
if [ -n "$warning" -a -n "$critical" ]
then
    if [ "$dn_avail" -le "$warning" -a "$dn_avail" -gt "$critical" ]
    then
        echo "WARNING - ${output} | ${perfdata}"
	exit $ST_WR
    elif [ "$dn_avail" -le "$critical" ]
    then
        echo "CRITICAL - ${output} | ${perfdata}"
	exit $ST_CR
    else
        echo "OK - ${output} | ${perfdata} "
	exit $ST_OK
    fi
else
    echo "OK - ${output} | ${perfdata}"
    exit $ST_OK
fi

Output example:

If everything went fine you should see output like the following:

user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin
OK - Datanodes up and running: 50, DFS total: 20147365 MB, DFS used: 0 MB (0%) | 'datanodes_available'=50 'dfs_total'=20147365 'dfs_used'=0
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -w 40 -c 30
OK - Datanodes up and running: 50, DFS total: 20147365 MB, DFS used: 0 MB (0%) | 'datanodes_available'=50 'dfs_total'=20147365 'dfs_used'=0
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -w 60 -c 40
WARNING - Datanodes up and running: 50, DFS total: 20147365 MB, DFS used: 0 MB (0%) | 'datanodes_available'=50 'dfs_total'=20147365 'dfs_used'=0
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -w 70-c 60
CRITICAL - Datanodes up and running: 50, DFS total: 20147365 MB, DFS used: 0 MB (0%) | 'datanodes_available'=50 'dfs_total'=20147365 'dfs_used'=0
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -w 20 -c 40
Please adjust your warning/critical thresholds. The warning must be higher than the critical level!
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -w 30
Please also set a critical value when you want to use warning/critical thresholds!
user@host: ~ $ ./check_hadoop-dfs.sh -s /var/nagios/home/bin -c 30
Please also set a warning value when you want to use warning/critical thresholds!

The License

Process Check Plugin for Nagios

Mike Adolphs — Sat, 30 May 2009 17:07:24 +0000

This is a sh-compliant check plugin for Nagios which enables you to monitor a specific process whether it’s running or not and to see how much cpu/memory it’s wasting. This can be accomplished with the check_procs plugin which is included in the official Nagios-Plugins package as well, but it doesn’t generate any performance output. At least that’s what I’ve experienced lately (I’d have written the plugin anyway for training purposes).

The script

No special needs, just the basic stuff. The script parses the output of ps aux in different ways. You need to provide a specific process via -p/–process, if warning/critical thresholds are wanted you may also choose between cpu or memory thresholds via -t/–target (default is memory).
As always, you may grab the plugin at MonitoringExchange or directly from here via SVN and the copy’n'paste way below.

user@host: ~ $ svn co svn://svn.matejunkie.com/nagios-plugins/stable/check_ps/ check_ps/

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
process="cron"
target="mem"
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to check a specific process via ps."
    echo "You may provide any string as an argument to match a specific"
    echo "process. Please note that the output could be distorted if the"
    echo "argument matches various processes, so please make sure to use"
    echo "unique strings to match a process."
    echo ""
    echo "$PROGNAME -p firefox [-w 10] [-c 20] [-t cpu]"
    echo ""
    echo "Options:"
    echo "  -p/--process)"
    echo "     You need to provide a string for which the ps output is then"
    echo "     then \"greped\"."
    echo "  -w/--warning)"
    echo "     Defines a warning level for a target which is explained"
    echo "     below. Default is: off"
    echo "  -c/--critical)"
    echo "     Defines a critical level for a target which is explained"
    echo "     below. Default is: off"
    echo "  -t/--target)"
    echo "     A target can be defined via -t. Choose between cpu and mem."
    echo "     Default is: mem"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        -help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --process|-p)
            process=$2
            shift
            ;;
        --target|-t)
            target=$2
            shift
            ;;
        --warning|-w)
            warning=$2
            shift
            ;;
        --critical|-c)
            critical=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
        esac
    shift
done
 
get_wcdiff() {
    if [ ! -z "$warning" -a ! -z "$critical" ]
    then
        wclvls=1
        if [ ${warning} -gt ${critical} ]
        then
            wcdiff=1
        fi
    elif [ ! -z "$warning" -a -z "$critical" ]
    then
        wcdiff=2
    elif [ -z "$warning" -a ! -z "$critical" ]
    then
        wcdiff=3
    fi
}
 
val_wcdiff() {
    if [ "$wcdiff" = 1 ]
    then
        echo "Please adjust your warning/critical thresholds. The warning \
must be lower than the critical level!"
        exit $ST_UK
    elif [ "$wcdiff" = 2 ]
    then
        echo "Please also set a critical value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    elif [ "$wcdiff" = 3 ]
    then
        echo "Please also set a warning value when you want to use \
warning/critical thresholds!"
        exit $ST_UK
    fi
}
 
get_vals() {
    process=`echo ${process} | sed 's/^.\|[a-z][A-Z] /\[&]/g'`
 
    tmp_output=`ps aux | grep "$process" | grep -v $0`
 
    if [ -z "$tmp_output" ]
    then
        echo "CRITICAL - Process is not running!"
        exit $ST_CR
    fi
 
    ps_user=`echo ${tmp_output} | awk '{print $1}'`
    ps_pid=`echo ${tmp_output} | awk '{print $2}' `
    ps_cpu=`echo ${tmp_output} | awk '{print $3}'`
    ps_mem=`echo ${tmp_output} | awk '{print $4}' `
    ps_start=`echo ${tmp_output} | awk '{print $9}' `
 
    tmp_ps_cputime=`echo ${tmp_output} | awk '{print $10}'`
    tmp_ps_cpuhours=`echo ${tmp_ps_cputime} | awk -F \: '{print $1}'`
    tmp_ps_cpumin=`echo ${tmp_ps_cputime} | awk -F \: '{print $2}'`
    ps_cputime=`echo "scale=0; (${tmp_ps_cpuhours} * 60) + \
${tmp_ps_cpumin}" | bc -l`
}
 
do_wccalc() {
    if [ -n "$warning" -a -n "$critical" ]
    then
        if [ "$target" = "cpu" ]
        then
            tmp_wc_target=`echo ${ps_cpu} | awk -F \. '{print $2}'`
            if [ "$tmp_wc_target" -ge 5 ]
            then
                wc_target=`echo ${ps_cpu} | awk -F \. '{print $1}'`
                wc_target=`expr ${wc_target} + 1`
            else
                wc_target=`echo ${ps_cpu} | awk -F \. '{print $1}'`
            fi
                elif [ "$target" = "mem" ]
                then
            tmp_wc_target=`echo ${ps_mem} | awk -F \. '{print $2}'`
            if [ "$tmp_wc_target" -ge 5 ]
            then
                wc_target=`echo ${ps_mem} | awk -F \. '{print $1}'`
                wc_target=`expr ${wc_target} + 1`
            else
                wc_target=`echo ${ps_mem} | awk -F \. '{print $1}'`
            fi
        fi
    fi
}
 
do_output() {
    process=`echo ${process} | sed 's/\[//g' |  sed 's/\]//g'`
    output="Process: ${process}, User: ${ps_user}, CPU: ${ps_cpu}%, \
RAM: ${ps_mem}%, Start: ${ps_start}, CPU Time: ${ps_cputime} min"
}
 
do_perfdata() {
    perfdata="'cpu'=${ps_cpu} 'memory'=${ps_mem} 'cputime'=${ps_cputime}"
}
 
# Here we go!
get_wcdiff
val_wcdiff
 
get_vals
do_wccalc
do_output
do_perfdata
 
if [ -n "$warning" -a -n "$critical" ]
then
    if [ "$wc_target" -ge "$warning" -a "$wc_target" -lt "$critical" ]
    then
        echo "WARNING - ${output} | ${perfdata}"
        exit $ST_WR
    elif [ "$wc_target" -ge "$critical" ]
    then
        echo "CRITICAL - ${output} | ${perfdata}"
        exit $ST_CR
    else
        echo "OK - ${output} | ${perfdata}"
        exit $ST_OK
    fi
else
    echo "OK - ${output} | ${perfdata}"
    exit $ST_OK
fi

PNP Template

#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

#   PNP Template for check_ps.sh
#   Author: Mike Adolphs (http://www.matejunkie.com/

$opt[1] = "--vertical-label \"percent\" -u 100 -l 0 -r --title \"CPU/Memory Usage for $hostname / $servicedesc\" ";
$opt[2] = "--vertical-label \"minutes\" -u 100 -l 0 -r --title \"cputime for $hostname / $servicedesc\" ";
 
$def[1] =  "DEF:cpu=$rrdfile:$DS[1]:AVERAGE " ;
$def[1] .=  "DEF:memory=$rrdfile:$DS[2]:AVERAGE " ;
$def[2] .=  "DEF:cputime=$rrdfile:$DS[3]:AVERAGE " ;
 
$def[1] .= "COMMENT:\"\\t\\t\\tLAST\\t\\t\\tAVERAGE\\t\\t\\tMAX\\n\" " ;
$def[2] .= "COMMENT:\"\\t\\t\\tLAST\\t\\t\\tAVERAGE\\t\\t\\tMAX\\n\" " ;
 
$def[1] .= "LINE2:cpu#E80C3E:\"CPU\\t\\t\" " ;
$def[1] .= "GPRINT:cpu:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:cpu:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:cpu:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "LINE2:memory#008000:\"Memory\\t\" " ;
$def[1] .= "GPRINT:memory:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:memory:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:memory:MAX:\"%6.2lf \\n\" " ;
 
$def[2] .= "AREA:cputime#E80C3E:\"CPUTime\\t\" " ;
$def[2] .= "GPRINT:cputime:LAST:\"%6.2lf min\\t\\t\" " ;
$def[2] .= "GPRINT:cputime:AVERAGE:\"%6.2lf min\\t\\t\" " ;
$def[2] .= "GPRINT:cputime:MAX:\"%6.2lf min\\n\" " ;

The License

Update 1.3: Bind check plugin for Nagios

Mike Adolphs — Wed, 20 May 2009 08:06:52 +0000

Thanks to Gareth’s feedback and advice I fixed a small, but nerving bug which raised an error when the plugin was called on a Debian based distribution.
The main problem here was, that the awk snippet failed because of an unescaped metacharacter when using mawk (Debian default awk alternative), but worked like a charm with gawk (Gentoo default). Apparently gawk is just more loose than mawk, it should’ve failed on both.

Anyway, this shouldn’t be a problem anymore. As always, the updated script is available at MonitoringExchange, or via svn. Please consider that the svn checkout is only possible via svnserve at the moment.

user@host: ~ $ svn co svn://svn.matejunkie.com/nagios-plugins/stable/check_bind/ check_bind/

CPU check plugin for Nagios

Mike Adolphs — Wed, 13 May 2009 10:48:13 +0000

I’ve just finished version 1.0 of a CPU check plugin which is sh-compliant (as always) and calculates utilization through Jiffies rather than using another third party tool like iostat or top. This means it should work on all Linux machines right out of the box without the need to install additional software.
Frankly speaking the performance of the script is slightly worse than the performance of most of the other CPU check plugins due to the sacrifice of parsing iostat or top output, but I nevertheless wanted to do this since calculating CPU utilization through Jiffies instead of parsing another tool’s output is somewhat different and interesting. I didn’t take a look into iostat’s or top’s source code yet, but I guess these tools doesn’t do it in another way than parsing /proc/stat as well. The difference in performance isn’t that huge anyway, a double-digit amount of milliseconds, but not more.

For getting along with the Jiffies, I crawled through the web, especially Paul Colby’s article about calculating CPU utilization through /proc/stat helped a lot!

The script

As always, the script is downloadable from NagiosExchange, via svn directly from here or the copy’n'paste way below. A proper PNP template is included as well.

user@host ~ $ svn co https://svn.matejunkie.com/svn/nagios-plugins/stable/check_cpu/ check_cpu/

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
interval=1
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to monitor CPU utilization. It makes"
    echo "use of /proc/stat and calculates it through Jiffies rather than"
    echo "using another frontend tool like iostat or top."
    echo "When using optional warning/critical thresholds all values except"
    echo "idle are aggregated and compared to the thresholds. There's"
    echo "currently no support for warning/critical thresholds for specific"
    echo "usage parameters."
    echo ""
    echo "$PROGNAME [-i/--interval] [-w/--warning] [-c/--critical]"
    echo ""
    echo "Options:"
    echo "  --interval|-i)"
    echo "    Defines the pause between the two times /proc/stat is being"
    echo "    parsed. Higher values could lead to more accurate result."
    echo "    Default is: 1 second"
    echo "  --warning|-w)"
    echo "    Sets a warning level for CPU user. Default is: off"
    echo "  --critical|-c)"
    echo "    Sets a critical level for CPU user. Default is: off"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        --help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --interval|-i)
            interval=$2
            shift
            ;;
        --warning|-w)
            warn=$2
            shift
            ;;
        --critical|-c)
            crit=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
    esac
    shift
done
 
val_wcdiff() {
    if [ ${warn} -gt ${crit} ]
    then
        wcdiff=1
    fi
}
 
get_cpuvals() {
    tmp1_cpu_user=`grep -m1 '^cpu' /proc/stat|awk '{print $2}'`
    tmp1_cpu_nice=`grep -m1 '^cpu' /proc/stat|awk '{print $3}'`
    tmp1_cpu_sys=`grep -m1 '^cpu' /proc/stat|awk '{print $4}'`
    tmp1_cpu_idle=`grep -m1 '^cpu' /proc/stat|awk '{print $5}'`
    tmp1_cpu_iowait=`grep -m1 '^cpu' /proc/stat|awk '{print $6}'`
    tmp1_cpu_irq=`grep -m1 '^cpu' /proc/stat|awk '{print $7}'`
    tmp1_cpu_softirq=`grep -m1 '^cpu' /proc/stat|awk '{print $8}'`
    tmp1_cpu_total=`expr $tmp1_cpu_user + $tmp1_cpu_nice + $tmp1_cpu_sys + \
$tmp1_cpu_idle + $tmp1_cpu_iowait + $tmp1_cpu_irq + $tmp1_cpu_softirq`
 
    sleep $interval
 
    tmp2_cpu_user=`grep -m1 '^cpu' /proc/stat|awk '{print $2}'`
    tmp2_cpu_nice=`grep -m1 '^cpu' /proc/stat|awk '{print $3}'`
    tmp2_cpu_sys=`grep -m1 '^cpu' /proc/stat|awk '{print $4}'`
    tmp2_cpu_idle=`grep -m1 '^cpu' /proc/stat|awk '{print $5}'`
    tmp2_cpu_iowait=`grep -m1 '^cpu' /proc/stat|awk '{print $6}'`
    tmp2_cpu_irq=`grep -m1 '^cpu' /proc/stat|awk '{print $7}'`
    tmp2_cpu_softirq=`grep -m1 '^cpu' /proc/stat|awk '{print $8}'`
    tmp2_cpu_total=`expr $tmp2_cpu_user + $tmp2_cpu_nice + $tmp2_cpu_sys + \
$tmp2_cpu_idle + $tmp2_cpu_iowait + $tmp2_cpu_irq + $tmp2_cpu_softirq`
 
    diff_cpu_user=`echo "${tmp2_cpu_user} - ${tmp1_cpu_user}" | bc -l`
    diff_cpu_nice=`echo "${tmp2_cpu_nice} - ${tmp1_cpu_nice}" | bc -l`
    diff_cpu_sys=`echo "${tmp2_cpu_sys} - ${tmp1_cpu_sys}" | bc -l`
    diff_cpu_idle=`echo "${tmp2_cpu_idle} - ${tmp1_cpu_idle}" | bc -l`
    diff_cpu_iowait=`echo "${tmp2_cpu_iowait} - ${tmp1_cpu_iowait}" | bc -l`
    diff_cpu_irq=`echo "${tmp2_cpu_irq} - ${tmp1_cpu_irq}" | bc -l`
    diff_cpu_softirq=`echo "${tmp2_cpu_softirq} - ${tmp1_cpu_softirq}" \
| bc -l`
    diff_cpu_total=`echo "${tmp2_cpu_total} - ${tmp1_cpu_total}" | bc -l`
 
    cpu_user=`echo "scale=2; (1000*${diff_cpu_user}/${diff_cpu_total}+5)/10" \
| bc -l | sed 's/^\./0./'`
    cpu_nice=`echo "scale=2; (1000*${diff_cpu_nice}/${diff_cpu_total}+5)/10" \
| bc -l | sed 's/^\./0./'`
    cpu_sys=`echo "scale=2; (1000*${diff_cpu_sys}/${diff_cpu_total}+5)/10" \
| bc -l | sed 's/^\./0./'`
    cpu_idle=`echo "scale=2; (1000*${diff_cpu_idle}/${diff_cpu_total}+5)/10" \
| bc -l | sed 's/^\./0./'`
    cpu_iowait=`echo "scale=2; (1000*${diff_cpu_iowait}/${diff_cpu_total}+5)\\
/10" | bc -l | sed 's/^\./0./'`
    cpu_irq=`echo "scale=2; (1000*${diff_cpu_irq}/${diff_cpu_total}+5)/10" \
| bc -l | sed 's/^\./0./'`
    cpu_softirq=`echo "scale=2; (1000*${diff_cpu_softirq}/${diff_cpu_total}\\
+5)/10" | bc -l | sed 's/^\./0./'`
    cpu_total=`echo "scale=2; (1000*${diff_cpu_total}/${diff_cpu_total}+5)\\
/10" | bc -l | sed 's/^\./0./'`
    cpu_usage=`echo "(${cpu_user}+${cpu_nice}+${cpu_sys}+${cpu_iowait}+\\
${cpu_irq}+${cpu_softirq})/1" | bc`
}
 
do_output() {
    output="user: ${cpu_user}, nice: ${cpu_nice}, sys: ${cpu_sys}, \
iowait: ${cpu_iowait}, irq: ${cpu_irq}, softirq: ${cpu_softirq} \
idle: ${cpu_idle}"
}
 
do_perfdata() {
    perfdata="'user'=${cpu_user} 'nice'=${cpu_nice} 'sys'=${cpu_sys} \
'softirq'=${cpu_softirq} 'iowait'=${cpu_iowait} 'irq'=${cpu_irq} \
'idle'=${cpu_idle}"
}
 
if [ -n "$warn" -a -n "$crit" ]
then
    val_wcdiff
    if [ "$wcdiff" = 1 ]
    then
        echo "Please adjust your warning/critical thresholds. The warning\\
must be lower than the critical level!"
        exit $ST_UK
    fi
fi
 
get_cpuvals
do_output
do_perfdata
 
if [ -n "$warn" -a -n "$crit" ]
then
    if [ "$cpu_usage" -ge "$warn" -a "$cpu_usage" -lt "$crit" ]
    then
        echo "WARNING - ${output} | ${perfdata}"
        exit $ST_WR
    elif [ "$cpu_usage" -ge "$crit" ]
    then
        echo "CRITICAL - ${output} | ${perfdata}"
        exit $ST_CR
    else
        echo "OK - ${output} | ${perfdata}"
        exit $ST_OK
    fi
else
    echo "OK - ${output} | ${perfdata}"
    exit $ST_OK
fi


 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

#   PNP Template for check_cpu.sh
#   Author: Mike Adolphs (http://www.matejunkie.com/

$opt[1] = "--vertical-label \"CPU [%]\" -u 100 -l 0 -r --title \"CPU Usage for $hostname / $servicedesc\" ";
 
$def[1] =  "DEF:used=$rrdfile:$DS[1]:AVERAGE " ;
$def[1] .=  "DEF:nice=$rrdfile:$DS[2]:AVERAGE " ;
$def[1] .=  "DEF:sys=$rrdfile:$DS[3]:AVERAGE " ;
$def[1] .=  "DEF:iowait=$rrdfile:$DS[4]:AVERAGE " ;
$def[1] .=  "DEF:irq=$rrdfile:$DS[5]:AVERAGE " ;
$def[1] .=  "DEF:softirq=$rrdfile:$DS[6]:AVERAGE " ;
$def[1] .=  "DEF:idle=$rrdfile:$DS[7]:AVERAGE " ;
 
$def[1] .= "COMMENT:\"\\t\\t\\tLAST\\t\\t\\tAVERAGE\\t\\t\\tMAX\\n\" " ;
 
$def[1] .= "AREA:used#E80C3E:\"user\\t\":STACK " ; 
$def[1] .= "GPRINT:used:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:used:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:used:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:nice#E8630C:\"nice\\t\":STACK " ; 
$def[1] .= "GPRINT:nice:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:nice:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:nice:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:sys#008000:\"sys\\t\t\":STACK " ;
$def[1] .= "GPRINT:sys:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:sys:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:sys:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:iowait#0CE84D:\"iowait\\t\":STACK " ;
$def[1] .= "GPRINT:iowait:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:iowait:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:iowait:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:irq#3E00FF:\"irq\\t\t\":STACK " ;
$def[1] .= "GPRINT:irq:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:irq:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:irq:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:softirq#1CC8E8:\"softirq\\t\":STACK " ;
$def[1] .= "GPRINT:softirq:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:softirq:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:softirq:MAX:\"%6.2lf \\n\" " ;
 
$def[1] .= "AREA:idle#FFFF00:\"idle\\t\":STACK " ; 
$def[1] .= "GPRINT:idle:LAST:\"%6.2lf %%\\t\\t\" " ;
$def[1] .= "GPRINT:idle:AVERAGE:\"%6.2lf \\t\\t\" " ;
$def[1] .= "GPRINT:idle:MAX:\"%6.2lf \\n\" " ;
?>

The License

As always this little script is ment to be sh-compliant and released under the terms of the GPL Version 2 only. Feel free to subscribe via rss to get updates on this one. More options will be added in the future.
I’ll probably add an option to make use of iostat or top and to calculate utilization per CPU soon so that you’re able to choose your own way.

Update 1.1: UDP Packet Loss Check Plugin for Nagios

Mike Adolphs — Mon, 11 May 2009 09:36:39 +0000

I’ve changed the small UDP packet loss script yesterday so that it’s now showing the total requests, requests sent to an unknown port and receive errors per check interval. That means, if you’d choose a check interval of 5 minutes, the script shows the amount of the previous mentioned for this time frame. It now creates two temporary files in a directory which you may define via -t/–path-tmp. The file is automatically deleted when it’s size got over 32Kb.
Feel free to grab it from MonitoringExchange (they’ve changed their name due to legal problems with Nagios Enterprises and because of Nagios’ new fork Icinga), via svn or copy’n'paste below.

user@host ~ $ svn co https://svn.matejunkie.com/svn/nagios-plugins/stable/check_udp-packet-loss/ check_udp-packet-loss/

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.1,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
path_tmp="/tmp"
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to monitor the UDP packet"
    echo "loss via netstat. It writes a temporary file to a configurable"
	echo "directory and compares it's values against the ones from the check"
	echo "before."
	echo "That means the actual result shows the amount of overall packets,"
	echo "received, receive errors and/or packets recieved to unknown port"
	echo "for the last check interval you've chosen."
    echo ""
    echo "$PROGNAME check_udp-packet-loss.sh"
    echo ""
    echo "Options:"
    echo "  --warning|-w)"
    echo "    Sets a warning level for packet receive errors. Default is:"
    echo "    off"
    echo "  --critical|-c)"
    echo "    Sets a critical level for packet receive errors. Default is:"
    echo "    off"
    echo "  --path-tmp|-t)"
    echo "    Sets the path for the temporary file. Default is: /tmp"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        --help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --path-tmp|-t)
            path_tmp=$2
            shift
            ;;
        --warning|-w)
            warn=$2
            shift
            ;;
        --critical|-c)
            crit=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
    esac
    shift
done
 
check_filesize() {
    if [ -f $path_tmp/udp-errors.log ]
    then
        size=`du $path_tmp/udp-errors.log | awk '{print $1}'`
	if [ $size -gt 32 ]
        then
            rm $path_tmp/udp-errors.log
            touch $path_tmp/udp-errors.log
        fi
    else
        touch $path_tmp/udp-errors.log
    fi   
}
 
get_packet_loss() {
	date_beg=`date`
	echo "+++ $date_beg +++" >> $path_tmp/udp-errors.log
    netstat -us|grep 'packets received' >> $path_tmp/udp-errors.log
    netstat -us|grep 'to unknown port' >> $path_tmp/udp-errors.log
    netstat -us|grep 'packet receive errors' >> $path_tmp/udp-errors.log
	date_end=`date`
	echo "--- $date_end ---" >> $path_tmp/udp-errors.log
	tac /tmp/udp-errors.log | awk '/---/{p=1} p{print} /+++/{p=0;if \
(count++==1) exit}' > $path_tmp/udp-errors.log.tmp
}
 
chk_packet_loss() {
	p_receive_1st=`tac $path_tmp/udp-errors.log.tmp | grep -m1 'packets \
received' | awk '{print $1}'`
	p_receive_2nd=`tac $path_tmp/udp-errors.log.tmp | grep 'packets \
received' | sort -r | grep -m1 'packets received' | awk '{print $1}'`
	p_unknown_1st=`tac $path_tmp/udp-errors.log.tmp | grep -m1 'to unknown \
port' | awk '{print $1}'`
	p_unknown_2nd=`tac $path_tmp/udp-errors.log.tmp | grep 'to unknown \
port' | sort -r | grep -m1 'to unknown port' | awk '{print $1}'`
	p_errors_1st=`tac $path_tmp/udp-errors.log.tmp | grep -m1 'receive \
errors' | awk '{print $1}'`
	p_errors_2nd=`tac $path_tmp/udp-errors.log.tmp | grep 'receive errors' \
| sort -r | grep -m1 'receive errors' | awk '{print $1}'`
 
	if [ ! $p_receive_2nd = 0 ]
	then
		p_receive=`expr $p_receive_2nd - $p_receive_1st`
	else
		p_receive=0
	fi
 
	if [ ! "$p_unknown_2nd" = 0 ]
	then
		p_unknown=`expr $p_unknown_2nd - $p_unknown_1st`
	else
		p_unknown=0
	fi
 
	if [ ! "$p_errors_2nd" = 0 ]
	then
		p_errors=`expr $p_errors_2nd - $p_errors_1st`
	else
		p_errors=0
	fi
}
 
val_wcdiff() {
    if [ ${warn} -gt ${crit} ]
    then
        wcdiff=1
    fi
}
 
check_filesize
get_packet_loss
chk_packet_loss
 
if [ -n "$warn" -a -n "$crit" ]
then
    val_wcdiff
    if [ "$wcdiff" = 1 ]
    then
        echo "The warning level should be lower than the critical level. \
Please adjust them."
        exit $ST_UK
    fi
fi
 
if [ -n "$warn" -a -n "$crit" ]
then
    if [ "$p_unknown" -ge "$warn" -a "$p_unknown" -lt "$crit" ]
    then
        echo "WARNING - $p_receive total packets received, $p_unknown \ 
went to unknown port, $p_errors receive errors! | \
'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_WR
    elif [ "$p_unknown" -ge "$crit" ]
    then
        echo "CRITICAL - $p_receive total packets received, $p_unknown \
went to unknown port, $p_errors receive errors! | \
'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_CR
    else
        echo "OK - $p_receive total packets received, $p_unknown \
went to unknown port, $p_errors receive errors! | \
'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_OK
    fi
else
    echo "OK - $p_receive total packets received, $p_unknown went to unknown \
port, $p_errors receive errors! | 'received'=$p_receive \
'unknown_port'=$p_unknown 'errors'=$p_errors"
    exit $ST_OK
fi

UDP Packet Loss Check Plugin for Nagios

Mike Adolphs — Tue, 05 May 2009 22:20:10 +0000

I just needed to rewrite a small check plugin I’ve written a few months ago to check possible UDP data loss on a host which serves an application that constantly has to process tens of thousand UDP packets at a time. Otherwise we just wouldn’t know in time when the application or the host’s network stack become somewhat limited because of UDP’s stateless nature and that’s what you better wanna know for sure if the transmitted data is valuable for the company.

The script

As always, the script is downloadable from NagiosExchange, via svn directly from here or the copy’n'paste way below.

user@host ~ $ svn co https://svn.matejunkie.com/svn/nagios-plugins/stable/check_udp-packet-loss/ check_udp-packet-loss/

It makes use of netstat to check for packets received to an unknown port or receive errors and to monitor the overall count of UDP packets received. This information is being put into a file in a temporary directory which is not being used at all for now, I must confess. The idea behind that is to rather monitor UDP packet receive errors over a period of time instead of its total count, but since there’s other heavy stuff going on at the moment, I just don’t have the time to finish this at the moment. I’ll do this within the next weeks.
The script checks by itselfs whether the file grows over 32K, then deletes and re-creates it. So no worries about broken hard drives here.

The term “packets to unknown port received” is explained as the following: “Displays the total number of UDP datagrams received for which there was no application at the destination port” which means that e.g. those will increase when you’d temporarily shutdown your application. “packet receive errors” will mostly occur when the host’s buffer is bursting. For additional information I’d suggest reading “UDP Buffer sizing” at 29west.com and UDP’s man page at kernel.org.

#!/bin/sh
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; either version 2 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
 
path_tmp="/tmp"
 
print_version() {
    echo "$VERSION $AUTHOR"
}
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to monitor the UDP packet loss via"
    echo "netstat."
    echo ""
    echo "$PROGNAME check_udp-packet-loss.sh"
    echo ""
    echo "Options:"
    echo "  --warning|-w)"
    echo "    Sets a warning level for packet receive errors. Default is:"
    echo "    off"
    echo "  --critical|-c)"
    echo "    Sets a critical level for packet receive errors. Default is:"
    echo "    off"
    echo "  --path-tmp|-t)"
    echo "    Sets the path for the temporary file. Default is: /tmp"
    exit $ST_UK
}
 
while test -n "$1"; do
    case "$1" in
        --help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --path-tmp|-t)
            path_tmp=$2
            shift
            ;;
        --warning|-w)
            warn=$2
            shift
            ;;
        --critical|-c)
            crit=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
    esac
    shift
done
 
check_filesize() {
    if [ -f $path_tmp/check_udp-packet-errors.log ]
    then
        size=`du $path_tmp/check_udp-packet-errors.log | awk '{print $1}'`
	if [ $size -gt 32 ]
        then
            rm $path_tmp/check_udp-packet-errors.log
            touch $path_tmp/check_udp-packet-errors.log
        fi
    else
        touch $path_tmp/check_udp-packet-errors.log
    fi
}
 
check_packet_loss() {
    p_receive=`netstat -us|grep 'packets received'|awk '{print $1}'`
    echo "Received: $p_receive" >> $path_tmp/check_udp-packet-errors.log
    p_unknown=`netstat -us|grep 'to unknown port'|awk '{print $1}'`
    echo "Unknown: $p_unknown" >> $path_tmp/check_udp-packet-errors.log
    p_errors=`netstat -us|grep 'packet receive errors'|awk '{print $1}'`
    echo "Errors: $p_errors" >> $path_tmp/check_udp-packet-errors.log
}
 
val_wcdiff() {
    if [ ${warn} -gt ${crit} ]
    then
        wcdiff=1
    fi
}
 
check_filesize
check_packet_loss
 
if [ -n "$warn" -a -n "$crit" ]
then
    val_wcdiff
    if [ "$wcdiff" = 1 ]
    then
        echo "The warning level should be lower than the critical level. Please adjust them."
        exit $ST_UK
    fi
fi
 
if [ -n "$warn" -a -n "$crit" ]
then
    if [ "$p_unknown" -ge "$warn" -a "$p_unknown" -lt "$crit" ]
    then
        echo "WARNING - $p_receive total packets received, $p_unknown went to unknown port, $p_errors receive errors! | 'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_WR
    elif [ "$p_unknown" -ge "$crit" ]
    then
        echo "CRITICAL - $p_receive total packets received, $p_unknown went to unknown port, $p_errors receive errors! | 'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_CR
    else
        echo "OK - $p_receive total packets received, $p_unknown went to unknown port, $p_errors receive errors! | 'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
        exit $ST_OK
    fi
else
    echo "OK - $p_receive total packets received, $p_unknown went to unknown port, $p_errors receive errors! | 'received'=$p_receive 'unknown_port'=$p_unknown 'errors'=$p_errors"
    exit $ST_OK
fi