Zabbix agent on {HOST.NAME} is unreachable for 5 minutes frequently

[rickychang]

I am new to Zabbix. I am using version 4.4.0.
The case is that I using the default template "Template App Zabbix Agent" which interval is 1 mins, and found that it always alert "Zabbix agent on {HOST.NAME} is unreachable for 5 minutes" after a day when i restart zabbix server.
Then I found that the item interval is increasing by itself and it is jump to around 1 hour interval, so that the trigger nodata(5m) is reached.
I found some solution but still not working. i.e. set 30 timeout on both server and agent.
The only way I found is to restart the zabbix server, but the case will appear after a day.
Is there any solution to resolve it?

[Atsushi]

4.4.0 is very buggy and should not be used. If you want to use Zabbix 4.4, try using the latest version 4.4.5 at the moment.

[the.monitor]

rickychang From my experience this looks like the agent lost connectivity with the proxy or server responsible for monitoring this.

If this has 1 min polling, it stands to reason that you lost communication with this machine.

There could be many reasons for this; maybe your NIC is saturated, Maybe the machine is too busy to respond.

The Zabbix agent check is NOT a simple check in the same way that ICMP is. It will check that the agent is responding to requests.

Check your graph for ICMP and other forms of monitoring during this same time period and i would be very surprised if some of them didn't have gaps in data.

EDIT: I've seen behavior like this on backup machines overnight when they're extremely busy.