Networknet archive Blog

Use Cacti to monitor your network latency

Originally published 33 comments

If you manage and monitor a network, you probably want to keep an eye on the latency of its links. This is especially useful for links using a dedicated internet connection and an IPSec VPN tunnel to the datacenter.

Latency is predictable, but for non-private IP VPN links without reserved bandwidth and QoS/CoS, monitoring it may help resolve problems or provide a basis for discussing high latency with the service provider. With dedicated Frame Relay, leased-line or IP VPN links, you can agree on latency values with the service provider and put SLAs in place.

Latency in a packet-switched network is measured either one-way (the time from the source sending a packet to the destination receiving it), or round-trip (the one-way latency from source to destination plus the one-way latency from the destination back to the source). Round-trip latency is more often quoted, because it can be measured from a single point. Note that round trip latency excludes the amount of time that a destination system spends processing the packet. Many software platforms provide a service called ping that can be used to measure round-trip latency. Ping performs no packet processing; it merely sends a response back when it receives a packet (i.e. performs a no-op), thus it is a relatively accurate way of measuring latency.

http://en.wikipedia.org/wiki/Latency_%28engineering%29

I use CactiEZ v0.3 to monitor ping latency. The hosts may or may not support SNMP. If the router is managed by the service provider, you probably will not have SNMP access, but this check does not need it.

Configure the latency graph

  1. Create a new device in the Cacti Management console. See the example for www.networknet.nl.

    image

  2. Set Availability to "Ping" and use "ICMP" as the ping method. Click Add and reopen the device.
  3. Locate Associated Graph Templates, choose "Unix-Ping Latency" in Graph Templates and click Add.

    image

  4. Click "*Create Graphs for this Host".
  5. Select the "Create: Unix - Ping Latency" graph template and click Create.
  6. Choose Red as the legend color and click Create. The graph is now created: + Created graph: www.networknet.nl - Ping Latency.
  7. Create a new Graph Tree and add the new network-latency graph. See my example.

    image

The result

image

33 comments

Comments from the original WordPress site, kept read-only.

  1. Hasmezi

    i want to monitor smtp server latency. the smtp service is up and running but i cannot ping the server's IP address. can u pls help me on the configuration to enable me to generate the latency graph for this particular server. i got few others smtp servers that i can generate the latency graph.

  5. Ali


    Hasmezi:

    i want to monitor smtp server latency. the smtp service is up and running but i cannot ping the server’s IP address. can u pls help me on the configuration to enable me to generate the latency graph for this particular server. i got few others smtp servers that i can generate the latency graph.

    maybe u can ping the nearest pingable network element to it. One that you are sure is on the same lan, otherwise u need to go with SNMP

  6. Sarmed

    I have already generated graphs, but nothing is being plotted. When I go to 'Device' I can see that cacti can ping the device. Dunno what's wrong?

  7. Eli

    If your graph isn't being plotted you may have the same problem I had. Not having ICMP as the ping type for step "b". I had UDP. Switching it to ICMP got it working.

  8. Sarmed

    Well, there was a bug in the cacti ping.pl script....i was doing grep for 'icmp_seq' to measure latency whereas my Debian host generates 'icmp_req' instead. That's why I was unable to plot graphs. After tweaking the script, latency graph worked like a charm. :)

  9. Olivier

    Sarmed,
    you re great ! now it works on my Debiantoo
    thx

    #open(PROCESS, "ping -c 1 $host | grep icmp_seq | grep time |");
    open(PROCESS, "ping -c 1 $host | grep icmp_req | grep time |");

  13. g3

    I'm using CentOS 6.3 but I keep getting the follow error when I run it I get an error:
    # open(PROCESS, “ping -c 1 $host | grep ‘icmp_[s|r]eq’ | grep time |”);
    bash: syntax error near unexpected token `PROCESS,'

  14. Sarmed

    Do a normal ping from your console. Check out whether you get a "icmp_req" or "icmp_seq" in the reply. Modify the perl script accordingly

    EITHER
    open(PROCESS, “ping -c 1 $host | grep icmp_req | grep time |”);
    OR
    open(PROCESS, “ping -c 1 $host | grep icmp_seq | grep time |”);

    Then run the script manually to check whether you find an error or not.

    I don't mean to do my publicity, but you could check out my blog if you want to.
    http://amar-linux.blogspot.com/2012/02/cacti-on-centos-6-setting-up-system.html

  15. g3

    Thanks Sarmed.

    This is what I get:

    64 bytes from ir1.fp.vip.sp2.yahoo.com (72.30.38.140): icmp_seq=2 ttl=250 time=101 ms
    64 bytes from ir1.fp.vip.sp2.yahoo.com (72.30.38.140): icmp_seq=3 ttl=250 time=104 ms

    [root@pdccentcacti01]# open(PROCESS, “ping -c 1 $host | grep icmp_seq | grep time |”);
    bash: syntax error near unexpected token `PROCESS,'

    Thanks for the link to your site also.

  16. Sarmed

    May I ask how you are running the script? I'm giving a sample run of one of my Debian servers, and the content of the script for your convenience.

    The platform is Debian, so only the part "icmp_req" is different from your Red Hat based script. The rest is the same.

    ############## SAMPLE RUN ##################
    root@dragonfly:/usr/share/cacti/site/scripts# perl ping.pl 192.168.1.3
    0.655

    ##########SCRIPT CONTENT############
    root@dragonfly:/usr/share/cacti/site/scripts# cat ping.pl
    #!/usr/bin/perl

    # take care for tcp:hostname or TCP:ip@
    $host = $ARGV[0];
    $host =~ s/tcp:/$1/gis;

    open(PROCESS, "ping -c 1 $host | grep icmp_req | grep time |");
    $ping = ;
    close(PROCESS);
    $ping =~ m/(.*time=)(.*) (ms|usec)/;

    if ($2 == "") {
    print "U"; # avoid cacti errors, but do not fake rrdtool stats
    }elsif ($3 eq "usec") {
    print $2/1000; # re-calculate in units of "ms"
    }else{
    print $2;
    }
    ###############################

  17. g3

    I'm in the / directory and I'm running the command from there (I'm very new at this)
    I used webmin to file the location of the file (nano /var/lib/cacti/scripts/ping.pl.)

    #!/usr/bin/perl

    # take care for tcp:hostname or TCP:ip@
    $host = $ARGV[0];
    $host =~ s/tcp:/$1/gis;

    # old linux version use "icmp_seq"
    # newer use "icmp_req" instead
    open(PROCESS, "ping -c 1 $host | grep 'icmp_req' | grep time |");
    $ping = ;
    close(PROCESS);
    $ping =~ m/(.*time=)(.*) (ms|usec)/;

    if ($2 == "") {
    print "U"; # avoid cacti errors, but do not fake rrdtool stats
    }elsif ($3 eq "usec") {
    print $2/1000; # re-calculate in units of "ms"
    }else{
    print $2;
    }

  19. Sarmed

    Correction:

    from the console, please post the output of the following-

    perl /var/lib/cacti/scripts/ping.pl #### ANY REACHABLE IP ####

  20. g3

    [root@pdccentcacti01]# perl /var/lib/cacti/scripts/ping.pl #### ANY REACHABLE IP ####
    Usage: ping [-LRUbdfnqrvVaA] [-c count] [-i interval] [-w deadline]
    [-p pattern] [-s packetsize] [-t ttl] [-I interface or address]
    [-M mtu discovery hint] [-S sndbuf]
    [ -T timestamp option ] [ -Q tos ] [hop1 ...] destination
    U[root@pdccentcacti01 ggraham]#

  22. Sarmed

    good to hear :)

    In the earlier post, i meant any reachable ip from your network e.g 4.2.2.2 by #### ANY REACHABLE IP ####

    but the problem is solved, and that's what counts. glad I could help.

  23. g3

    Sarmed,

    What should the output look like after: perl /var/lib/cacti/scripts/ping.pl 192.168.1.1

  26. Ashish Gupta

    Hi Team;

    Added Window/Linux host without SNMP information @ Cacti but not getting any Latency information on cacti graph

  27. Ashish Gupta

    Team;

    Added Window/Linux host without SNMP information @ Cacti but not getting any Latency information on cacti graph

    Ashish Gupta

  28. sundarkanna

    Hi all,

    My graph was working to monitor ICMP report. suddenly graph not generating and not working.
    please guide anyone that what is template name to make live the same.

    Thanks,
    sundar

  29. vladimir

    Me puedes enviar por favor el template Unix - Ping Latency; lo borre por accidente y no lo encuntro en internet.

    Gracias,

  30. khan

    my all interface working fine but ping latency is not working .can any boday help on this please .. i have migrate my machine from old to new .please help how ping latency graph will work.

  33. Paul Jeyasingh

    I have multiple interfaces to my cacti server. And have different subnet IP's to each interfaces. I want to initiate ping to different destinatoin thru different soruce IP's. Is it possible?

    For Eg:- eth0 to google.com and eth1 to yahoo.com