Search Exchange

Search All Sites

Nagios Live Webinars

Let our experts show you how Nagios can help your organization.

Contact Us

Phone: 1-888-NAGIOS-1
Email: sales@nagios.com

Login

Remember Me

check_iftraffic64

Rating
23 votes
Favoured:
4
Current Version
.77
Last Release Date
2018-03-27
Compatible With
  • Nagios 1.x
  • Nagios 2.x
  • Nagios 3.x
  • Nagios 4.x
  • Nagios XI
License
GPL
Hits
78318
Files:
FileDescription
check_iftraffic64.plCheck plugin
check_net_status.phpPNP4Nagios template.
Network Monitoring Software - Download Nagios XI
Log Management Software - Nagios Log Server - Download
Netflow Analysis Software - Nagios Network Analyzer - Download
check_iftraffic64
Bug and feature update to check_iftraffic3. This is a 64 (and 32) bit Nagvis compatible SNMP iftraffic check. Renamed to highlight the change from 32 to 64 bit as the default counters used. Tested with Windows 2003/08/12, Linux (RedHat), and Cisco devices.
Alert: If you are having problems with results sometimes returning higher than 100% IN or OUT utilization try setting the check command to forced 32 or 64 bit mode with the --force switch. Several reviewers have posted this problem but I did not understand what was happening until Rafael van den Berg helped me see the cause, thanks Rafael. I don't know if this will resolve everyone’s issue but if Rafael is right then this should address most if not all of the over 100% utilization scenarios. By default, check_iftraffic64 auto detects between 32 and 64 bit counters, trying 64 bit first. In cases where the 64 bit SNMP query doesn't always respond (but does sometimes) you will get mixed results (64 bit results last check 32 bit this check) and your output will be nonsensical (over 100%). Setting the check to force 32 or 64 will prevent this from affecting your results negatively in those cases. See examples below for how to use the force switch.

Check_iftraffic64 is an updated version of (check_iftraffic3) and has been changed to use 64 bit SNMP counters but can still handle 32 bit. Some math issues found in check_iftraffic3 were resolved as well.


Version .77: There are no functionality changes with this release. I simply fixed some code comments (spelling errors, removed spurious comments, etc.)

I also uploaded a PNP4Nagios template for the plugin. It provides the following features. Put this file in the /usr/local/pnp4nagios/share/templates/ directory. See uploaded examples.

1. It converts the performance data from the Nagios plugin standard (Bytes) to bits, however, the graph is percent based, bits are listed in the text of the legend not on the graph itself.
2. Peak - Prints the timestamp of when your peak traffic during the graph period happened.
3. Rem Peak Cap - Calculates remaining capacity or headroom (how much bandwidth is left between your peak traffic and your critical threshold).
4. Rem Ave Cap - Calculates remaining capacity or headroom (how much bandwidth is left between your average traffic and your critical threshold).
5. Total Bytes - Displays the total amount of data that has crossed the interface during the graphed period.
6. Min over Thresh - Displays how much time during the graphed period traffic was above the critical threshold.
7. Times over Thresh - Counts and displays how many times the critical threshold was crossed during the graphed period.

Version .76: Thanks to Mark Rittinghaus for pointing out and recommending solutions to a couple of bugs in the perfdata. These have been addressed in this release. This release changes the units for the in_ave/out_ave perfdata from bits (if you specify bits on the command line) to Bytes and changes inAbsolut/outAbsolut from Bytes to Continuous. Both of these changes may disrupt your perfdata graphing if you use PNP4Nagios (RRD). It does not affect the in_ave_pct/out_ave_pct perfdata output.

Version .75:
Thanks to Philip Ho and Fai Wong (and subsequently others) for pointing out that the inbandwidth and outbandwidth perfdata needed to be a fixed unit for rrd based processing of perfdata (PNP4Nagios). This has been fixed based on the switches used to either bits or bytes. The other bug which has plagued this plugin for a while is the nefarious and random 'Return code of 25 is out of bounds' result. This has also been fixed, again thanks to the guidance of Philip Ho .

Version .74:
Thanks to the help of Rogerio Tomassoni de Araujo Junior we now have SNMPv3 support! The check can use three v3 levels (NoauthNoPriv, authNoPriv, and authpriv). You can also specify the Auth method (MD5 or SHA) and or the encryption method (DES, AES, or 3DES). Thanks Rogerio!

Version .73:
Fixed a couple of bugs, they are listed in the comments at the top of the script.

Version .7:
The check script now has an 64|32 bit auto detection, you can also force 64 or 32 bit on the cmd line. Added logic to handle invalid interface speed results (thanks Mathieu GRZYBEK).

Version .6:
I've removed text from the output that was not necessary. Thanks Nicholas Coulin for pointing it out.


Version .5:
Significant features in this check include the following:

1. Capable of handling 32 and 64 (default) bit SNMP counters
2. Auto detects max bandwidth (unless specified)
3. Supports asymetric links, i.e. Ability to specify different BW limits (in and out) other than what is automatically detected. Good for Internet connections that have different up and down speeds
4. NagVis weathermap line compatible perfdata output
5. Returns CRITICAL status when the interface is down.


examples:

# Simple 64 bit check of interface used as the primary host interface (based on IP address of host1)
check_iftraffic64.pl -H host1 -C sneaky

# 32 bit mode check of interface index 5 in bits/s with 100m bandwidth limit
check_iftraffic64.pl -H host1 -C sneaky -i 5 -B -b 100 -u m --32bit

# Check of interface using address 192.168.1.1 in bits/s running 50m down (in) and 10m up (out)
check_iftraffic64.pl -H host1 -C sneaky -A 192.168.1.1 -B -I 50 -O 10 -u m
Reviews (17)
Hello,

have a problem here with this script and snmpd service 5.7.3 on Ubuntu18.04
We ask for traffic on remote host on eth0 by:
check_iftraffic64.pl -H 10.100.4.193 -i eth0
Result is:CRITICAL: Could not match eth0

Reason is line "my $resp=$response->{$key};" in sub fetch_ifdescr
On a working machine $resp contains lo, eth0, eth1
On invalid machine with new snmpd this contains also interface/controllerinformation, so script does not find the eth0 match.
###
my $resp=$response->{$key};
$resp =~ s/x00//;
print "resp: $resp
";
###
This is response from snmpd on Ubuntu18.04
resp: Intel Corporation Ethernet Controller 10-Gigabit X540-AT2
resp: Intel Corporation 82574L Gigabit Network Connection
resp: Intel Corporation Ethernet Controller 10-Gigabit X540-AT2
resp: Intel Corporation 82574L Gigabit Network Connection
resp: lo

So i have some assumptions:

- this Script losts some of values of response, unfortunately iam not a perl-programmer
- snmpd on machine is buggy
- snmpd on machine has wrong configuration. (We use minimized /etc/snmp/snmpd.conf with content "rocommunity public 10.100.4.206" successful on many systems)

What do you mean? Is there something know to this case?

Thanks,
Hans
/usr/bin/perl /usr/local/nagios/libexec/check_iftraffic64.pl -H firewall_IP -i 7 -I 15 -O 15 -b 15 -u m -f -L -w 80 -c 90 -v 3 --username --authpassword --authprotocol MD5 --privpassword --privprotocol AES

On executing above command from CLI on server i'm getting output as
"CRITICAL - OUT bandwidth (109.13%) too high|inUsage=32.14%;80;90 outUsage=109.13%;80;90 inBandwidth=602577.37B outBandwidth=2046264.78B inAbsolut=736506469220c outAbsolut=382869632218c"

BUt on web interface it is showing output as "(No output on stdout) stderr:"
bysalluste, November 13, 2017
hi all
I use check_iftraffic64 with centreon 3.4. snmpV3 I had a problem. I had a response from the plugins, the check after i had no output fromthe plugins, the check after my response is ok. it's the same if i do it from the shell. could you help me or give me some advice to solve my problem thanks a lot
Hi,

I tested this plugin and it did not work correctly out of the box on the command line because it did output nothing at all (but web ui said "OK"). Looking into the code the problem was on line 140 where it says
my $TRAFFIC_FILE = "/usr/local/nagios/libexec/traffic/";

This directory did not exist on my system; this caused the plugin to output nothing. After adding this directory, it worked. Maybe add a test if directory exists and throw an error if not?

I miss two features though, whose availability would easily push this plugin to a five star plugin(!)
- error packet counter and corresponding warn/crit tresholds
- automatic port speed detection

These two are very important IMHO:
- error counter
If the error counters grows, you could be alarmed; thanks to performance data, you would even have some long time trending, being able to research when errors increased e.g.

- auto-detect port speed
If the network admin decides to change a port speed from 100M to 1G, the plugin settings for this specific port will become incorrect. Auto-detection could help here.

If you start the plugin with debug level 4 you can see that port speed detection _would_ already work. The command

/usr/lib/nagios/plugins/check_iftraffic64.pl -H 192.168.123.123 -C not-so-public -i 1 --debug 4

outputs (shortend):

OID's:
IfOperStatus: 1.3.6.1.2.1.2.2.1.8.1
IfSpeed: 1.3.6.1.2.1.31.1.1.1.15.1
IfSpeed32: 1.3.6.1.2.1.2.2.1.5.1
RESULTS:
Operational Status: 1
Interface Speed (64bit): 100 Mbits
Interface Speed (32bit): 100000000 bits

So there is already a port speed detection, but not used in any way - maybe it is unreliable?

Since I cannot re-review a plugin here, I will mark it with 5 stars, hoping the author might add these two features in the future.
bySolidd, June 6, 2016
I'm using centreon, and when I run the plugin, no output appear, except if I run it in debug mode.

It's because the tmp plugin path has changed since.

I solved the problem, by making sure the temp directory exist & have right (or just change it in the script ;))

Thanks for the author !
byGeoHolz, October 12, 2015
Hello,

Can you add the possibility to have the output with only one unit ?
Actually, the result can be in B KB MB GB
For PNP4Nagios we need measure with always the same unit

Thanks
Owner's reply

Hi GeoHolz,

Yes, the fix was just posted in version .75. Thanks for pointing out the deficiency.

-greg

This check works great on most of my devices, but now iam testing it on a Cisco ASA5505 and there is a problem:

On full interface saturation the traffic is shown around 200% of the interface. I am downloading a testfile with ~5Mb, but the check reports ~10Mb.

Its a 50/10MBit wan-interface and this is my check:
check_iftraffic64.pl # -H 192.168.123.123 -C public -i "Adaptive Security Appliance 'outside' interface" -I 50 -O 10 -u m
Owner's reply

Hi Unic,

If you are specifying bandwidth limits with the -I and -O switches then the plugin is going to use those parameters in calculations. In that case if your getting more than 100% throughput the only thing I can think of is that the circuit is faster that listed.

byabeckman702, April 27, 2015
1 of 1 people found this review helpful
Overall this works quite well. I do have an issue on a few nodes that it reports high numbers like stated in 2 of the previous reviews. I would like to get this fixed as this is used primarily for the perfdata for nagvis. With this issue I hear alarms throughout the day. CRITICAL - IN bandwidth (35711048.32%) too high|inUsage=35711048.32%;85;98 outUsage=69084009.66%;85;98 inBandwidth=892.78GB outBandwidth=1727.10GB inAbsolut=4463881040328B outAbsolut=8635501207011B
Owner's reply

Hi,

I am not sure of all cases that would cause this. There are a few I am aware of though.

1. We had a NAS device that returned the 32 bit counter value for both 32 bit and 64 bit SNMP queries. This took us a while to figure and we were surprised to see them doing this. This will give you the types of results you are seeing, though I imaging (and hope) that this is a rare situation.

2. Another case is when two instances of Nagios are writing to the temp file for an interface messing with the calculations of the other check.

There may be other causes but these are the two that come to mind. You can try the latest version of the plugin there may be something else that was fixed there that would address this as well.

byogirard, March 12, 2015
1 of 1 people found this review helpful
Very good Job. But I've a problem with the result:
./check_iftraffic64.pl -H 192.168.16.254 -C public -i wan2 -b 20 -u m
CRITICAL - IN bandwidth (35711048.32%) too high|inUsage=35711048.32%;85;98 outUsage=69084009.66%;85;98 inBandwidth=892.78GB outBandwidth=1727.10GB inAbsolut=4463881040328B outAbsolut=8635501207011B

Do you know why ?
Thx a lot
Owner's reply

Hi ogirard,

If you add '--force' to your check command it should resolve your issue. See the ALERT section of the check description for an explanation of the issue.

cheers,

-greg

bymorube, January 13, 2015
Really usefull on my network. Is this the right configuration to control the gigabitethernet interface Bandwidth load?

/check_iftraffic64.pl -H DEVICE_IP -C COMMUNITY -i GigabitEthernet0/48 -I 1000 -O 1000 -u g -w 85 -c 95

Thanks!!
bynotic, December 13, 2014
1 of 1 people found this review helpful
Hi,
this plugin works very well, but i am getting this error on few devices periodically.

Traffic_gi0/2;CRITICAL;SOFT;1;CRITICAL - IN bandwidth (81985529197.52%) too high
Owner's reply

Hi notic,

Use the '--force' switch in your command. See the ALERT section of the description above for details.

-greg

Have spent most of the day trying to get various bandwidth monitors working for a linux machine, and this is excellent.

Just one small bug. The following line is executed without debug mode being requested:

print "$ifdescr = $key / $snmpkey \n"; #debug

For me, this results in the first line of the status text to be:

eth0 = 1.3.6.1.2.1.2.2.1.2.2 / 2

and the following actual lines aren't displayed in the host service list (Service Status Details For Host), only when drilling down to the specific service (Service State Information).

I've just commented out the line above, and the script no longer outputs the interface name and OID.
Owner's reply

Hi,

Thanks for pointing this out. The bug has been fixed, thanks to Gerhard M. who also reported this bug.

-greg

bybwen, September 11, 2013
-b said it require argument
-b 100 display the help, same for --bandwidth 100 or 10 or 1 or 1000
Owner's reply

The -b switch does require an argument, it also requires the -u switch which specifies the units. For example you would do some thing like: '-b 100 -u m' for an interface speed of 100mb.

It works excellent on Cisco switches, but it doesn't work on Brocade switches. Any plans to support Brocade SAN switches?

Regards
Domen
Owner's reply

Hi domensetar,

The check does not intentionally discriminate against Brocade. It is using standard SNMP API's in perl to query the end devices. I'm not sure why they don't work. We don't use Brocade devices here, nor do i have any to test with. If you are still trying to resolve this you could send me an example of the errors you are getting, I may be able to figure it out.

-greg

Plugin some time return error
"Illegal division by zero at ./check_iftraffic64.pl line 427" when runing in console
"Return code of 25 is out of bounds" when used in nagios
Owner's reply

I believe the "Illegal division" issue has been fixed. The other one however, 'Return code of 25..." has not been resolved. I have been unable to track it down but suspect it has something to do with the nagios host not able to open any more network connections for the script. We see that one in our environment but it is intermittent. I don't understand the relationship between the perl interpreter and the network sockets ona machine, is there a limit to how many a process can open? I may be way off on this.

by_bk201, November 1, 2012
2 of 3 people found this review helpful
Awesome plugin, thanks!

Also if you are having issues with "error code 13 out of bounds" you probably tested it with root user on command line first, which creates a 'traffic' folder in the 'libexec' folder with root ownership, change owner and or group to your nagios user to fix this.
bymeilon, October 21, 2012
It's good, but the SNMPv3 support should be added soon
Owner's reply

I agree that SNMPv3 should be added to make the check more complete. Currently this is not a requirement by my employer, until that changes I won't have any cycles to offer in that regard. If you want to add that functionality, I would upload it here.

Regards,

-greg