Alert if process is down for more than 1 minute

[apalacheno]

Howdy,

monitoring if a process is running is easy with a trigger like:

{Template_Services:proc.num[ddclient - slee].last(0)}=0

But how can I tell Zabbix to only fire the trigger if the process is down for more than a minute?

Use case: Some processes get routinely restarted. That shouldn't fire a trigger each time. Similiarily, if a routine update gets installed, the process is shutdown for a few seconds during the update; that shouldn't lead to a notification either.

I already tried with this:

{Template_Services:proc.num[ddclient - slee].min(120)}<1

but this sets off an alarm as soon as the process is down.

Soo... how can I tell Zabbix to alert only if a process is down for more than 1 minute?

Cheers,

Robert

[akbar415]

Howdy,

monitoring if a process is running is easy with a trigger like:

{Template_Services:proc.num[ddclient - slee].last(0)}=0

But how can I tell Zabbix to only fire the trigger if the process is down for more than a minute?

Use case: Some processes get routinely restarted. That shouldn't fire a trigger each time. Similiarily, if a routine update gets installed, the process is shutdown for a few seconds during the update; that shouldn't lead to a notification either.

I already tried with this:

{Template_Services:proc.num[ddclient - slee].min(120)}<1

but this sets off an alarm as soon as the process is down.

Soo... how can I tell Zabbix to alert only if a process is down for more than 1 minute?

Cheers,

Robert

{Template_Services:proc.num[ddclient - slee].last(2m)}=0

Or with histeresys

({TRIGGER.VALUE}=0 and
{Template_Services:proc.num[ddclient - slee].max(2m)}=0)
or
({TRIGGER.VALUE}=1 and
{Template_Services:proc.num[ddclient - slee].min(1m)}<1)

Trigger will fired if in the last 2 minutes the value was 0 (no process running), and will remaining until the value was more than 0 for at least 1 min.


Sorry for the bad englsih

[apalacheno]

Hi akbar,

thank you very much for your answer! It led me directly to the right configuration.

Checking if a process is down for X minutes can be done by two ways:

{Template_Services:proc.num[ddclient - slee].max(Xm)}<1

or

{Template_Services:proc.num[ddclient - slee].sum(Xm)}=0

They both seem equal to me; however for readability I prefer the latter.

The last(X) function cannot be used, since according to this sheet last() is always equal to last(#1). I can confirm this by own tests.

Cheers,

Robert

[apalacheno]

Just to follow up, I've now settled for the following trigger expression:

{Template_Services:proc.num[ddclient - slee].max(60)}=0

Result: If the process 'ddclient - slee' is down for more than 60 seconds, the trigger will get fired.