Host restart trigger last for days

[mellis]

We have started to have the trigger for the host restart last over 24hours in the dashboard. We do a monthly OS update and reboot on our windows servers. How can we get it to go away after 15mins

[tim.mooney]

Can you show us what the trigger is? I'm interested in 3 things

1. The "Problem Expression"
2. Whether there is a "Recovery Expression". You can click on the "Recovery Expression" box next to "OK Event Generation" to display whether there is one or not.
3. What the setting is for "OK Event closes"

If you provide that information, it should be enough to determine what the problem is.

[mellis]

1. The "Problem Expression"
   1. {Template OS Windows - No Ping:system.uptime.change(0)}<0
2. Whether there is a "Recovery Expression". You can click on the "Recovery Expression" box next to "OK Event Generation" to display whether there is one or not.
   1. No recovery Expression
3. What the setting is for "OK Event closes"
   1. All Problems

[tim.mooney]

Thanks for providing that information.

According to the function documentation , change() isn't listed as taking any arguments, so change(0) may not be right. My guess is it's either silently ignored or the documentation is incomplete and change() does actually take a time parameter.

The intent of the trigger seems fairly straightforward: by using change() and comparing <0, the trigger is essentially saying "if the previous value for system.uptime is larger than the current value for system.uptime, there's a problem".

Now that I know that change() is part of the trigger, I have one follow-up question: How often is this item (system.uptime) being collected for your hosts? Specifically what are the values for the item's "Update interval" and is there a "Custom Interval" set?

[mellis]

We check this status every 15 mins

[tim.mooney]

I'm not sure what would be causing the problem you're seeing. You've verified that you're collecting the item data on a regular basis, and the trigger seems correct.

If I were having this problem, my next step would probably be to increase the zabbix_serverd debug logging and then reboot a test Windows client that has this template item & trigger applied to it, to see if I can spot anything in the debug output that would show why the trigger isn't clearing.