Ad Widget

Collapse

Agent ping going haywire

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Eudemonia
    Junior Member
    • Mar 2015
    • 4

    #1

    Agent ping going haywire

    A bit of background: our store has 7 Linux servers and 3 Windows. We installed Zabbix agents on each of them, and they're all being monitored just fine. Set up an email system for warning triggers, all seemed to be going well. I wanted to change the default "zabbix agent unreachable for 5 minutes" trigger to 1 minute intervals like so:

    {Template App Zabbix Agent:agent.ping.nodata(1m)}=1

    Doing this caused the system to freak out. I started getting a continuous stream of emails from three of the ten servers, 2 Linux and 1 Windows, saying "PROBLEM: Zabbix agent on [hostname] is unreachable for 1 minute", followed up by a resolution "OK" email. I was able to fix the issue by changing the Template App Zabbix Agent ping update interval from 60 seconds to 30.

    Could this be some issue with the item's update interval being out of sync with the alert trigger? Is there documentation about this issue?
  • aib
    Senior Member
    • Jan 2014
    • 1615

    #2
    Well, it's better to have Update interval shorter than interval of inactivity in Trigger.
    Or Trigger will try to fire the alarm every time when it has no data, because data is not updated yet.

    And keep in mind - update interval is not fixed. When you configure it to 60 second - it is 60tish, i.e. any value from 58 to 63 seconds, not exactly 60!
    Sincerely yours,
    Aleksey

    Comment

    • Eudemonia
      Junior Member
      • Mar 2015
      • 4

      #3
      Shouldn't the Zabbix server still have last stored information even after 60 seconds have passed? Or does it just wipe the stored value after the given update interval?

      Comment

      Working...