Zabbix 7 SNMP Timeouts

[zendx]

Hello! I decided to test Zabbix 7 (beta3).
Unfortunately, there are problems with the absence of discovery for some hosts, where two devices of the same model and firmware version can be discovered differently (one successfully, the other not).
I noticed that Zabbix closes the connection prematurely, with a timeout of 5 seconds, which is not enough to receive a response from the network device.
Is it possible to increase this timeout?
The connection is dropped (While responses from the device 172.17.49.5 are still coming in) right at the "SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds" log message in the zabbix log (

641615:20240426:121433.952 End of zbx_ipc_async_socket_recv():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000123 sec, getting values]'
641615:20240426:121433.952 In get_values()
641615:20240426:121433.952 In zbx_dc_config_get_poller_items() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_items():0
641615:20240426:121433.952 In zbx_dc_config_get_poller_nextcheck() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_nextcheck():1714122878
641615:20240426:121433.952 End of get_values():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000130 sec, idle 5 sec]'
641615:20240426:121433.952 In zbx_ipc_async_socket_recv() timeout:5
641643:20240426:121436.188 SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds

[LukeAB93UK]

Hello! I decided to test Zabbix 7 (beta3).
Unfortunately, there are problems with the absence of discovery for some hosts, where two devices of the same model and firmware version can be discovered differently (one successfully, the other not).
I noticed that Zabbix closes the connection prematurely, with a timeout of 5 seconds, which is not enough to receive a response from the network device.
Is it possible to increase this timeout?
The connection is dropped (While responses from the device 172.17.49.5 are still coming in) right at the "SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds" log message in the zabbix log (

641615:20240426:121433.952 End of zbx_ipc_async_socket_recv():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000123 sec, getting values]'
641615:20240426:121433.952 In get_values()
641615:20240426:121433.952 In zbx_dc_config_get_poller_items() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_items():0
641615:20240426:121433.952 In zbx_dc_config_get_poller_nextcheck() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_nextcheck():1714122878
641615:20240426:121433.952 End of get_values():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000130 sec, idle 5 sec]'
641615:20240426:121433.952 In zbx_ipc_async_socket_recv() timeout:5
641643:20240426:121436.188 SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds

Hello,

In the Zabbix 7 beta did you configure timeouts? I believe this is a new feature.

5 What's new in Zabbix 7.0.0

1 General (zabbix.com)

Those are the instructions on configuring timeouts.

[LukeAB93UK]

Hello! I decided to test Zabbix 7 (beta3).
Unfortunately, there are problems with the absence of discovery for some hosts, where two devices of the same model and firmware version can be discovered differently (one successfully, the other not).
I noticed that Zabbix closes the connection prematurely, with a timeout of 5 seconds, which is not enough to receive a response from the network device.
Is it possible to increase this timeout?
The connection is dropped (While responses from the device 172.17.49.5 are still coming in) right at the "SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds" log message in the zabbix log (

641615:20240426:121433.952 End of zbx_ipc_async_socket_recv():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000123 sec, getting values]'
641615:20240426:121433.952 In get_values()
641615:20240426:121433.952 In zbx_dc_config_get_poller_items() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_items():0
641615:20240426:121433.952 In zbx_dc_config_get_poller_nextcheck() poller_type:0
641615:20240426:121433.952 End of zbx_dc_config_get_poller_nextcheck():1714122878
641615:20240426:121433.952 End of get_values():0
641615:20240426:121433.952 zbx_setproctitle() title:'poller #4 [got 0 values in 0.000130 sec, idle 5 sec]'
641615:20240426:121433.952 In zbx_ipc_async_socket_recv() timeout:5
641643:20240426:121436.188 SNMP agent item "net.if.walk" on host "testhost" failed: first network error, wait for 15 seconds

Hello,

In the Zabbix 7 beta did you configure timeouts? I believe this is a new feature.

5 What's new in Zabbix 7.0.0

1 General (zabbix.com)

Those are the instructions on configuring timeouts.

[troffasky]

Weird. Updated to Zabbix 7 2 weeks ago and I only just noticed this issue after applying the new Zabbix "Mikrotik by SNMP" template to a few APs.

EDIT: it's specifically the "walk" SNMP items that cause this behaviour. All the other items on the host work fine but then because the walk takes to long, it looks like Zabbix just gives up on SNMP for the whole host.


3774575:20240622:225115.656 SNMP agent item "sensor.temp.walk" on host "miaap" failed: another network error, wait for 15 seconds
3774542:20240622:225130.553 resuming SNMP agent checks on host "miaap": connection restored
3774575:20240622:225130.648 SNMP agent item "sensor.cpu.temp.walk" on host "miaap" failed: first network error, wait for 15 seconds
3774575:20240622:225145.658 SNMP agent item "sensor.cpu.temp.walk" on host "miaap" failed: another network error, wait for 15 seconds
3774542:20240622:225200.602 resuming SNMP agent checks on host "miaap": connection restored
3774575:20240622:225200.650 SNMP agent item "sensor.cpu.temp.walk" on host "miaap" failed: first network error, wait for 15 seconds
3774575:20240622:225215.658 SNMP agent item "sensor.cpu.temp.walk" on host "miaap" failed: another network error, wait for 15 seconds
3774542:20240622:225230.639 resuming SNMP agent checks on host "miaap": connection restored
3774575:20240622:225234.647 SNMP agent item "net.if.wireless.walk" on host "miaap" failed: first network error, wait for 15 seconds
3774575:20240622:225249.648 SNMP agent item "sensor.temp.walk" on host "miaap" failed: another network error, wait for 15 seconds
3774575:20240622:225304.659 SNMP agent item "sensor.cpu.temp.walk" on host "miaap" failed: another network error, wait for 15 seconds
3774575:20240622:225319.657 temporarily disabling SNMP agent checks on host "miaap": interface unavailable

[mikecel79]

I am seeing similar behavior when trying to monitor a FortiGate and running Zabbix 7.01. The SNMP walk items fail and then it just seems to give ups monitoring SNMP at all.