Multi Router Traffic Grapher
MRTG (Multi Router Traffic Grapher) monitors SNMP network devices and generates graphs showing how much traffic has passed through each interface.
Server setup
This document assumes that you already have a Apache and net-snmp working and configured properly.
Install mrtgAUR and perl-net-snmp.
- create an mrtg user
# useradd -d /srv/http/mrtg mrtg
- create the user home directory and change the owner ship to the user
# mkdir /srv/http/mrtg/ # chown mrtg:mrtg /srv/http/mrtg
Apache configuration
As far as the Apache configuration is concerned, we simply need to add an alias which is directed to the HTML files locations :
The configuration should look like this :
Alias /mrtg /srv/http/mrtg/html/ <Directory "/srv/http/mrtg/html/"> AllowOverride None Options None DirectoryIndex index.html Order allow,deny Allow from all </Directory>
MRTG setup
There are many ways to configure the mrtg for your local server. Here, the easiest way to expand the application for other server and network appliances is described if needed.
The following should all be performed as the mrtg user we created.
- create an HTML directory to hold the png files and the index.html file
# mkdir /srv/http/mrtg/html
now we will begin dealing with the application scripts first we will create a basic mrtg.cfg file
- The following script call will scan our localhost for its interfaces and create for us the relevant configuration for each interface; public is the community name set for the local SNMP access:
# cfgmaker --output=/srv/http/mrtg/mrtg.cfg --ifref=name --ifref=descr --global "WorkDir: /srv/http/mrtg" public@localhost
- the mrtg.cfg files contains all the server interfaces. we do not need the "lo" interface so we are going to delete it and edit the global configuration
mrtg.cfg Global configuration
remove the lines that are irrelevant to the interface and add the fallowing lines at the top:
### Global configuration ### LoadMIBs: /usr/share/snmp/mibs/UCD-SNMP-MIB.txt EnableIPv6: no HtmlDir: /srv/http/mrtg/html ImageDir: /srv/http/mrtg/html LogDir: /srv/http/mrtg ThreshDir: /srv/http/mrtg RunAsDaemon: Yes Interval: 5 Refresh: 600
the global configuration lines mean :
- 1) to load the Linux MIB in mrtg
- 2) to enable/disable IPv6
- 3) HTML home directory
- 4) the png files home directory
- 5) the log dir files locations
- 6) the Thresh folder
- 7) whether or not we want to run the application as a daemon , in this case : yes
- 8) the daemon interval (minimum 5 min)
- 9) the interval to refresh the HTML files
Resource monitoring
Now that we have the global configuration set we need to add the resources and devices we want to monitor.
in this tutorial we are going to monitor:
- 1)CPU
- 2)Memory Usage
- 3)swap
- 4)Number of Processes
- 5)Total TCP Established Connections
- 6)Users Count
- 7)the server mount points
- 8)the server interfaces
CPU monitoring
for monitoring the CPU we need to add the next lines :
Target[localhost.cpu]:ssCpuRawUser.0&ssCpuRawUser.0:[email protected] + ssCpuRawSystem.0&ssCpuRawSystem.0:[email protected] + ssCpuRawNice.0&ssCpuRawNice.0:[email protected] RouterUptime[localhost.cpu]: [email protected] MaxBytes[localhost.cpu]: 100 Title[localhost.cpu]: CPU Load
PageTop[localhost.cpu]: <h1>Active CPU Load %
Unscaled[localhost.cpu]: ymwd ShortLegend[localhost.cpu]: % YLegend[localhost.cpu]: CPU Utilization Legend1[localhost.cpu]: Active CPU in % (Load) Legend2[localhost.cpu]: Legend3[localhost.cpu]: Legend4[localhost.cpu]: LegendI[localhost.cpu]: Active LegendO[localhost.cpu]: Options[localhost.cpu]: growright,nopercent
Memory usage
to monitor the memory usage we need to add the next lines :
# get memory Usage Target[localhost.memtotal]: ( .1.3.6.1.4.1.2021.4.5.0&.1.3.6.1.4.1.2021.4.5.0:public@localhost ) - ( .1.3.6.1.4.1.2021.4.6.0&.1.3.6.1.4.1.2021.4.6.0:public@localhost )
PageTop[localhost.memtotal]: <h1>Memory Usage
Options[localhost.memtotal]: nopercent,growright,gauge Title[localhost.memtotal]: Memory Usage MaxBytes[localhost.memtotal]: 100000000 kMG[localhost.memtotal]: k,M,G,T,P,X YLegend[localhost.memtotal]: bytes ShortLegend[localhost.memtotal]: bytes LegendI[localhost.memtotal]: Memory Usage: LegendO[localhost.memtotal]: Legend1[localhost.memtotal]: Memory Usage, not including swap, in bytes Colours[localhost.memtotal]: Blue#1000ff, Black#000000, Gray#CCCCCC, Yellow#FFFF00
Swap usage
for swap usage add the following lines :
# get swap memory Target[localhost.swap]:( .1.3.6.1.4.1.2021.4.3.0&.1.3.6.1.4.1.2021.4.3.0:public@localhost ) - ( .1.3.6.1.4.1.2021.4.4.0&.1.3.6.1.4.1.2021.4.4.0:public@localhost)
PageTop[localhost.swap]: <h1>Swap Usage
Options[localhost.swap]: nopercent,growright,gauge,noinfo Title[localhost.swap]: Swap Usage MaxBytes[localhost.swap]: 100000000 kMG[localhost.swap]: k,M,G,T,P,X YLegend[localhost.swap]: bytes ShortLegend[localhost.swap]: bytes LegendI[localhost.swap]: Swap Usage: LegendO[localhost.swap]: Legend1[localhost.swap]: Swap memory avail, in bytes Colours[localhost.swap]: Blue#1000ff,Violet#ff00ff,Black#000000, Gray#CCCCCC
in the title section some calculation are made. MRTG knows to calculate the values given from the OID
Number of processes
for getting the number of processes running we are doing some unique here
# get number of processes running Target[localhost.procs]: `/usr/local/mrtg/linux_porc.pl` Title[localhost.procs]: Process Statistics
PageTop[localhost.procs]: <h1>Process Statistics
MaxBytes[localhost.procs]: 10000 YLegend[localhost.procs]: Processes LegendI[localhost.procs]: Blocked Processes: LegendO[localhost.procs]: Run Queue: Legend1[localhost.procs]: Number of Blocked Processes Legend2[localhost.procs]: Number of Processes in Run Queue Legend3[localhost.procs]: Maximal Blocked Processes Legend4[localhost.procs]: Maximal Processes in Run Queue Options[localhost.procs]: growright, integer, nopercent, gauge
as we can see here we are calling the command linux_proc.pl that was written in Perl and returns an Integer which presence the number of processes.
the content of the command is :
#!/usr/bin/perl open(COMD,"ps -ef | wc -l|"); $num = <COMD>; close(COMD); print int($num);
Established connections
in order to get a graph about established connections we are doing the way as the privies section :
# get number of established connections Target[localhost.estconn]: `/usr/local/mrtg/linux_estconn.pl` Title[localhost.estconn]: Established connections
PageTop[localhost.estconn]: <h1>Established connections
MaxBytes[localhost.estconn]: 100000 YLegend[localhost.estconn]: Established connections LegendI[localhost.estconn]: Established connections: LegendO[localhost.estconn]: Number of Established connections: Options[localhost.estconn]: growright, integer, nopercent, gauge Colours[localhost.estconn]: Red#FF0000,Blue#0066CC,Black#000000, White#FFFFFF
the content of the file linux_estconn.pl is :
#!/usr/bin/perl open(COMD,"ss -an | grep ESTABLISHED | wc -l|"); $num = <COMD>; close(COMD); print int($num);
Users count
for the users count once again we are using a Perl script to create an integer output
for the mrtg configuration we need to add:
# get number of current users Target[localhost.users]: `/usr/local/mrtg/linux_users.pl` Title[localhost.users]: logged in users
PageTop[localhost.users]: <h1>number of users
MaxBytes[localhost.users]: 100000 YLegend[localhost.users]: users count Legend0[localhost.users]: logged in users count: Options[localhost.users]: growright, integer, nopercent, gauge Colours[localhost.users]: Red#FF0000,White#FFFFFF,Blue#0066CC,Black#000000
The linux_users.pl file content is:
#!/usr/bin/perl open(COMD,"w | grep -v load | grep -v USER | wc -l|"); $num = <COMD>; close(COMD); print int($num);
Monitor mount points
in order to monitor mount points we first need to make sure that SNMP is sending us the relevant information to check the mount point OID we need first to see all the mount points by the command :
snmpwalk -v 2c -c public localhost mount
this will display all of the server mount points and there mount location.
to monitor the mount point we want we need to take the last octet from the result and add it to the next 2 OID's
.1.3.6.1.4.1.2021.9.1.8. .1.3.6.1.4.1.2021.9.1.6.
so the mrtg.cfg section for the root FS will look like this:
# monitor root FS Target[localhost.rootfs]: .1.3.6.1.4.1.2021.9.1.8.1&.1.3.6.1.4.1.2021.9.1.6.1:public@localhost
PageTop[localhost.rootfs]: <h1>Root FS Usage
Options[localhost.rootfs]: nopercent,growright,gauge,noinfo Title[localhost.rootfs]: Root FS Usage MaxBytes[localhost.rootfs]: 100000000 YLegend[localhost.rootfs]: Giga bytes ShortLegend[localhost.rootfs]: bytes LegendI[localhost.rootfs]: Root FS Usage: Colours[localhost.rootfs]: Yellow#FFFF00, White#FFFFFF, Gray#CCCCCC, Blue#1000ff
Index page
the index page is automatically generated when we run the "indexmaker ./mrtg.cfg > index.html" command.
Startup script
If you want the MRTG daemon to start at boot add the next startup script:
vi /etc/rc.d/mrtg
#!/bin/bash . /etc/rc.conf . /etc/rc.d/functions LANG=C USER=mrtg MRTG=/usr/bin/mrtg MRTGCFG=/srv/http/mrtg/mrtg.cfg daemon_name=mrtg Start() { stat_busy "starting the MRTG Daemon" su - ${USER} -c "env LANG=${LANG} ${MRTG} ${MRTGCFG} > /dev/null" RETVAL=$?; if [[ $RETVAL -eq 0 ]]; then add_daemon $daemon_name stat_done else stat_fail exit 1 fi } Stop() { stat_busy "Stopping the MRTG Daemon" PID=`ps -ef | grep mrtg.cfg | grep -v grep | awk '{print $2}'` if [[ ! -z ${PID} ]]; then kill ${PID} RETVAL=$?; if [[ $RETVAL -eq 0 ]]; then rm_daemon $daemon_name stat_done else stat_fail exit 1 fi fi } case "$1" in start) Start; ;; stop) Stop; ;; restart) Stop; Start; ;; *) echo "Usage: mrtg {start|stop|restart}"; ;; esac