False triggers - count not working properly

[klabarre]

I admit, there's a chance I do not understand how this works. But believe me when I say I've read the manual and many sites via Google.

Looking at the following trigger expression I believe the status of the server would need to be "2" three times before the trigger "triggers". Correct? This trigger is triggering after a single value of "2".

{server1:status.max(#3)}=2

Here's another:

This trigger should only trigger if http_perf returns "0" three times in a row, correct? However, it's triggering after a single value of "0".

{server1:http_perf.max(#3)}=0

Someone please explain what I'm missing. I'm coming from Nagios and we're now losing sleep due to the false alarms.

Kevin

[untergeek]

With the max(#n) polls, I've tended to use greater-than and less-than rather than a straight-up equals. In principle your triggers look perfect, so I don't know what to say. Try making the second one <1.

As far as using "status" goes, I've never had anything but trouble from that check. It has been problematic, so we simply deleted it and use other checks to ensure the server is up, e.g. tcp,10050 for zabbix agent listening, etc.

[DSon]

If you haven't already discovered the answer...

Max is the maximum value for a number of checks (or period of time, if you omit the # symbol).. for example:

{server1.max(#3)}='x' - this means the max value = 'x' for the last 3 values
{server1.max(300)}='x' - this means the max value = 'x' for the last 300 secs

You need to use Count instead.. for example:

{server1.count(#3)}='x' - the number of values = 'x' for the last 3 values

Hope this helps,
Danny.