ExtremeCloud IQ- Site Engine & Extreme Management Center

  • 1.  XMC monitoring philosophy

    Posted 11-30-2018 14:39
    I'm currently evaluating and as far as I see XMC is relaying on data what are sent by devices either with syslog or snmp trap. I know that XMC is snmp polling the device for status and interfaces. Is it also possible to snmp poll a specific OID and raise an alarm if the returned value does not match the definition?

    Regards

    Dany


  • 2.  RE: XMC monitoring philosophy

    Posted 12-03-2018 12:00
    I am pretts sure this is possible, but indeed currently i am not able to tell you how ...


  • 3.  RE: XMC monitoring philosophy

    Posted 12-03-2018 12:10
    I I checked all config paramaters in XMC and also the documentation, but I could not find it.


  • 4.  RE: XMC monitoring philosophy

    Posted 12-03-2018 19:40
    I am very sorry but i have to recall my comment.

    Monitoring a special oid and having an alarm seems currently NOT possible with XMC V8.1.x.

    I hope in near future version.


  • 5.  RE: XMC monitoring philosophy

    Posted 12-04-2018 08:03
    Hello Dany,

    I agree with M. Nees. This function does not exist at the moment.
    But you can create a small script to check the oid and create an syslog entry if you receive a value not matching your definition.

    That syslog event can trigger an alarm, too.

    Regards
    Stephan


  • 6.  RE: XMC monitoring philosophy

    Posted 12-04-2018 20:42
    Hello Dany,

    here a short working Python example, running directly on an XMC Ubuntu installation (not on Windows!!!) without additional Python modules.
    You will find a corresponding entry in you syslog on XMC, have a look.


    code:
    import commands

    #Change OID as you need. Here we request the SysContact.
    commando = "/usr/bin/snmpget -v3 -u username -l authPriv -a SHA -x AES -X password -A password switchip 1.3.6.1.2.1.1.4.0"

    ans= commands.getoutput(commando)

    print(ans)

    alarm = commands.getoutput("logger" + " " + ans)

    print(alarm)


    Have fun.

    Regards
    Stephan


  • 7.  RE: XMC monitoring philosophy

    Posted 12-05-2018 12:34
    Hey thanks for your answers and the script template!
    Our current monitoring system is mainly based on snmp status alarming. I'm considering to change towards syslog and trap alarming, because the effort to maintain all the monitors for different devices is huge. I'm only worried that syslog or traps might get lost in transport. For example you have a power supply failure and the device is sending a syslog or trap (udp) and the massge gets lost, you will probably never notice that. What are your opinions and experiences?


  • 8.  RE: XMC monitoring philosophy

    Posted 12-05-2018 12:49
    Hi Dany,

    normally errors which stay longer like fan or PSU failures will be reported via trap serveral times within a fixed interval.
    So if you loose one message the next will come few seconds or minutes later.

    Regards