Cisco IP SLA monitoring

                                

        
        

            
            

                

    

    
Cisco IP SLA monitoring


Description


This plugin checks for the status of IP SLA probes configured on Cisco routers.


IP SLAs can be configured as follows on routers (tested on IOS 12.4T):


! Ping probe example
! (target must be another Cisco box configured as IP SLA responder)
ip sla 1
icmp-jitter 10.10.10.10
tag PING
frequency 300


! Jitter probe example
! (target must be another Cisco box configured as IP SLA responder)
ip sla 2
udp-jitter 10.10.10.10 5000
tag JITTER
frequency 300
 
! VoIP G729a probe example
! (target must be another Cisco box configured as IP SLA responder)
ip sla 3
udp-jitter 10.10.10.10 5000 codec g729a
tag VOIP
frequency 300


! DNS probe example
ip sla 4
dns www.google.com name-server 8.8.8.8
tag DNS
frequency 300


! HTTP probe example


ip sla 5
http get http://www.google.com/ name-server 8.8.8.8 cache disable
tag HTTP
frequency 300


Notes


You may also want to configure some thresholds to make the probe fail
when it takes too long for instance. This is quite important because this
Nagios plugin will not accept warning and critical boundaries. These should
be configured on the IP SLA itself.
This plugin only checks the return status of such IP SLAs and it succeeds
only when IP SLAs succeed on the remote router. In other words, this plugin
checks if the configured IP SLA is met or not.


Modes


Single SLA mode


In this mode you specify which IP SLA you want to check (identified by Tag). The plugin
will then connect to the router using SNMP (v2c for now), make sure that there's an IP SLA configured
that corresponds to the supplied Tag, retrieve some probe specific information and report its
status back to Nagios (including probe specific performance data).


Pros & Cons:
      + granular monitoring of IP SLAs
      + probe specific status information & perf data
      - less scalable; possibly many IP SLA services per router
 
Example 1:


$ ./check_cisco_ipsla.pl -H 10.10.10.10 -c public -i DNS


IP SLA 4 # DNS ( dns probe to www.google.com ) status: OK | RTT=41;Query=www.google.com
 
Example 2:


$ ./check_cisco_ipsla.pl -H 10.10.10.10 -c public -i HTTP


IP SLA 5 # HTTP ( http probe to http://www.google.com/ ) status: OK | DNS=2;TCP=3;HTTP=4;Total=9;Query=http://www.google.com/


Multi SLA mode


In this mode you specify the keyword "ALL" as IP SLA tag (i.e. "-i ALL" on command line). The plugin
will then connect to the router using SNMP (v2c for now), list all configured IP SLAs, retrieve their status,
and report all that information back to Nagios. If only one IP SLA fails, the whole Nagios service will fail.
Also, the plugin will report the completion time of all IP SLAs configured in the performace data section.
In this mode the plugin generates a multi-line output.


Pros & Cons:


     - no granular monitoring of IP SLAs
     - no probe specific status information & perf data
     + scalable; only one IP SLA service per router


Example 3:


$ ./check_cisco_ipsla.pl -H 10.10.10.10 -c public -i ALL


IP SLA status: OK | Total-Time=469
IP SLA 1 # PING_London ( icmp-jitter ) status: OK
IP SLA 2 # JITTER_Tokyo ( udp-jitter ) status: OK
IP SLA 3 # VOIP_NewYork ( udp-jitter g729a ) status: OK
IP SLA 4 # DNS_Paris ( dns ) status: OK
IP SLA 5 # HTTP_Google ( http ) status: OK | PING_London=18
JITTER_Tokyo=250
VOIP_NewYork=106
DNS_Paris=32
HTTP_Google=81
Version 0.1:


Changelog


New features:


    

  • First public release
    • 

  • Single SLA mode + Multi SLA mode
    • 

  • Support for HTTP probe
    • 

  • Support for DNS probe
    • 

  • Support for VOIP/g729a probe
    • 

  • Support for ICMP probe
    • 

  • Support for Jitter probe
    • 



Todo list:


    

  • P1:Add support for all remaining probe types
    • 

  • P2:Better error reporting
    • 

  • P2:Clean Perl programming
    • 

  • P3:Evaluate perf data from plugin (warning+critical thresholds)
    • 

  • P4:Dynamically set & run probes without prior router config
    • 



Known bugs:
 - BUG0001: supplying a wrong SNMP community string makes the plugin always succeed!
 - BUG0002: in multi SLA mode, an empty line is displayed after the 1st one