Escalation how?

[safl]

Hello!

Does anybody have a reasonable notification system running with zabbix?
Thing is that the tech crew is getting flooded with messages.
I've got actions on triggers to be auto generated for all triggers.
But i would very much like to limit the messages being sent, eg. only sending the mails when a trigger has been true for 10min and not when it immediatly changes state.

Thing is that checks such as CPU can for a short period of time flip between the trigger value and non-trigger value. this results in severel emails being generated with messages like: CPU on (on) CPU on (off) CPU on (on) CPU on (off). Please share your experiences.

I've tried using escalations but i quite simply don't know how they work.

regards,

Simon

[Nate Bell]

You might want to take a good long look at what the Techs really needs to be notified about. If you can't afford to scale back the number of triggers, then try modifying the triggers to be less sensitive. You could also increase the amount of time between checks for your items, and make use of the .last() and .prev() values. If the last and prev values are both above/below your threshold, then the trigger goes off.

A feature that Zabbix doesn't have, but seems like it would help you, (and others including myself), would be one that lets you fetch a value from a specific time period. That way you could compare values over a period of time to see if you're just getting a spike, or a trend of unwanted behavior.

If all else fails, you could craft some custom alert scripts. Perhaps one that writes all the non-emergecy warnings to a file, and then a cron job sends the file in an email to your tech people. That way they get all their information, but in a single email.

Nate