"SNMP down" connection alert does not go up

[Krakenn]

Hello everyone,
I'm in charge of setting up Zabbix so that it reports the operating status of 4 devices (FW, Printer, Router and Switch) on 60 sites.
Hosts are set up, SNMP info is sent.
I'd like an alert to appear in the dashboard when SNMP info stops coming back (disconnection or devices switched off or malfunctioning).
By default, an SNMP disconnection doesn't appear as a problem.
I've created an Item with this information:
Type: internal Zabbix
Key: zabbix[host,snmp,available]
Information type: numeric (unsigned)
update interval: 1m
Storage period: 90d

Then I linked a trigger to it (I confirm the link: the trigger appears below the element)
Severity: Disaster
Expression: last(/IMP-XXX/zabbix[host,snmp,available],#300)=0
(the expression has been built with the expression builder and tested: it returns 0 when not connected to SNMP).

The logs show a record of 0s every minute for the device in question.
The problem does NOT appear in the dashboard, nor in the device problems.
The trigger does appear in the list of triggers for the device to be monitored, BUT, the status is noted as "UNKNOWN" and an Information in red:
Cannot evaluate function last(/IMP-XXX/zabbix[host,snmp,available],#300): not enough data.
There is more than enough data available in the log:
Extract from the element monitoring the printer's SNMP connection:
01/09/2023 11:22:26
0
01/09/2023 11:21:26
0
01/09/2023 11:20:26
0
01/09/2023 11:19:26
0
01/09/2023 11:18:26
0
01/09/2023 11:17:26
0
01/09/2023 11:16:26
0
01/09/2023 11:15:26
0
01/09/2023 11:14:26
0
01/09/2023 11:13:26
0
01/09/2023 11:12:26
0
01/09/2023 11:11:26
0
01/09/2023 11:10:26
0
01/09/2023 11:09:26
0
01/09/2023 11:08:26
0
01/09/2023 11:07:26
0
01/09/2023 11:06:26
0
01/09/2023 11:05:26
0
01/09/2023 11:04:26
0


Thanks for any constructive help in bringing this information up in the interface issues.

Translated with www.DeepL.com/Translator (free version)

[SanMonitor]

Hello everyone,
I'm in charge of setting up Zabbix so that it reports the operating status of 4 devices (FW, Printer, Router and Switch) on 60 sites.
Hosts are set up, SNMP info is sent.
I'd like an alert to appear in the dashboard when SNMP info stops coming back (disconnection or devices switched off or malfunctioning).
By default, an SNMP disconnection doesn't appear as a problem.
I've created an Item with this information:
Type: internal Zabbix
Key: zabbix[host,snmp,available]
Information type: numeric (unsigned)
update interval: 1m
Storage period: 90d

Then I linked a trigger to it (I confirm the link: the trigger appears below the element)
Severity: Disaster
Expression: last(/IMP-XXX/zabbix[host,snmp,available],#300)=0
(the expression has been built with the expression builder and tested: it returns 0 when not connected to SNMP).

The logs show a record of 0s every minute for the device in question.
The problem does NOT appear in the dashboard, nor in the device problems.
The trigger does appear in the list of triggers for the device to be monitored, BUT, the status is noted as "UNKNOWN" and an Information in red:
Cannot evaluate function last(/IMP-XXX/zabbix[host,snmp,available],#300): not enough data.
There is more than enough data available in the log:
Extract from the element monitoring the printer's SNMP connection:
01/09/2023 11:22:26
0
01/09/2023 11:21:26
0
01/09/2023 11:20:26
0
01/09/2023 11:19:26
0
01/09/2023 11:18:26
0
01/09/2023 11:17:26
0
01/09/2023 11:16:26
0
01/09/2023 11:15:26
0
01/09/2023 11:14:26
0
01/09/2023 11:13:26
0
01/09/2023 11:12:26
0
01/09/2023 11:11:26
0
01/09/2023 11:10:26
0
01/09/2023 11:09:26
0
01/09/2023 11:08:26
0
01/09/2023 11:07:26
0
01/09/2023 11:06:26
0
01/09/2023 11:05:26
0
01/09/2023 11:04:26
0


Thanks for any constructive help in bringing this information up in the interface issues.

Translated with www.DeepL.com/Translator (free version)

With these settings the SNMP has to be unavailable for 5 hours (1 item every minute * 300), are you sure it's been logging that long? Just checking since it seems quite a long time to wait before you trigger the SNMP unavailable.

[cyber]

With these settings the SNMP has to be unavailable for 5 hours (1 item every minute * 300), are you sure it's been logging that long? Just checking since it seems quite a long time to wait before you trigger the SNMP unavailable.

NO,... last(/host/item,#300) check for 300th value back from NOW... not for 300 values... There is a difference how # works for last and for example min/max... So if your trigger says "Cannot evaluate function last(/IMP-XXX/zabbix[host,snmp,available],#300): not enough data.", then most probably you do not have 300 checks worth of data yet...

https://www.zabbix.com/documentation/6.0/en/manual/appendix/functions/history

last (/host/key,<#num<:time shift>>)
#num (optional) - the Nth most recent value

I don't see explanation, what topic starter actually wanted to achieve with #300? 300 sec i.e. 5 minutes? Use max(/host/item,5m)<1 or max(/host/item,#5)<1

[dunnsland]

I am having exactly the same issue and reading what was said I cannot figure out how to resolve this.



This shows that the device has gone offline, not reachable by ping (SNMP unavailable)
Looking at the dashboard I see no indication that the AP is down or offline.



What am I doing wrong and how do I get the device to show on the dashboard as offline.

Thanks
Lawrence