Ad Widget

Collapse

Agent is unreachable OK alert without corresponding Problem alert

Collapse
X
Collapse
 
  • Page of 1
  • Time
  • Show
Clear All
new posts
Previous template Next
  • michaill
    Junior Member
    • May 2014
    • 18
    #1

    Agent is unreachable OK alert without corresponding Problem alert

    Zabbix 2.4.4 on Oracle Linux 6.

    Zabbix periodically raises "Agent is unreachable OK" alert without ever having sent the problem alert for that trigger. Looking at data for agent.ping shows that it is always 1, every minute, including specifically during that alert time. Very strange.

    This happens only for a few Windows servers (2012) out of around 100 Windows servers that are monitored. All those servers are in same rack if that helps. The alerts come in around once every day or two per server.

    How's that possible?

    EDIT:

    Looking at values closer, every time an OK alert was sent, the value for previous minute for agent.ping was missing.
    However, the trigger is :agent.ping.nodata(6m)}=1, and only one value was missing. And in any case, there was never a PROBLEM alert for that trigger.

    Any suggestions appreciated,
    Mike
    Last edited by michaill; 22-06-2017, 05:42.
    Tags: None
    • batchenr
      Senior Member
      • Sep 2016
      • 440
      #2
      Originally posted by michaill
      Zabbix 2.4.4 on Oracle Linux 6.

      Zabbix periodically raises "Agent is unreachable OK" alert without ever having sent the problem alert for that trigger. Looking at data for agent.ping shows that it is always 1, every minute, including specifically during that alert time. Very strange.

      This happens only for a few Windows servers (2012) out of around 100 Windows servers that are monitored. All those servers are in same rack if that helps. The alerts come in around once every day or two per server.

      How's that possible?

      EDIT:

      Looking at values closer, every time an OK alert was sent, the value for previous minute for agent.ping was missing.
      However, the trigger is :agent.ping.nodata(6m)}=1, and only one value was missing. And in any case, there was never a PROBLEM alert for that trigger.

      Any suggestions appreciated,
      Mike

      i have this to sometimes when our machines backing up at night - if you have it in these time maybe tell the agent not to check at this time.

      check this windows server agent.log at the time it sends the ok and zabbix_Server.log i bet ull see something like lost connection
      and what is your zabbix server version ?

        Comment

        Previous template Next
        Working...