Ad Widget

Collapse

Host Status Delay Long Time!

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • RadicaLover
    Junior Member
    • Sep 2005
    • 7

    #1

    Host Status Delay Long Time!

    I'm monitoring a host with the status and network load.

    So I have a trigger "Host is down" (using item icmpping (simple check)) and others triggers "Hight load in","hight load out" (using snmp (ifinoctects, ifoutoctects).

    All works fine with the host up.

    But when i shut down the host, looking at page alarms history if the alarm "host is down" comes first (i think i dont have control about the sequence) it works fine, but if comes first one of the "load alarm"(snmp) the host status change imediatly to unreacheable and the trigger "host is down" dont appear. "Host is down" only appears a lot of time later (20 min or more), and i want that trrigger imediatly!

    I think this is beacause the snmp retrys a lot of times with time delays and because that dont do the pingcheck, or maybe beacause after the host staus is unracheable no more pigchecks are done.

    What can i do so that the trigger "host is down" appears imediatly??

    Can you help me plis??

    Last edited by RadicaLover; 16-09-2005, 18:26.
  • James Wells
    Senior Member
    • Jun 2005
    • 664

    #2
    Greetings,

    This is all going to depend on your check periodicity. Generally, due to the time it takes for SNMP to be performed, you will want to set it to a longer delay than things like icmpping. When you do this, you will generally fail the status check sooner than you will the SNMP checks. Additionally, you generally want the SNMP checks to be dependent on either the status or the icmpping checks. That way, they will not alarm when the other checks are in a failed state.
    Unofficial Zabbix Developer

    Comment

    • RadicaLover
      Junior Member
      • Sep 2005
      • 7

      #3
      TANK YOU JAMES!

      All SNMP checks are dependents from icmpping. This works fine if icmpping check fails first, but the problem is that not always is icmpping to fail first!

      The check periodicity isnt relevant for this!
      I have the SNMP checks with 30sec. interval and the icmpping with 8sec.
      But as the SNMP checks are much more than (one) icmpping, it could append that one SNMP check occurs first than the icmpping (and is not so rare) and the problem stays.

      Comment

      Working...