Troubleshooting Inconsistent Zabbix Agent Active Check Availability Across Hosts

[firdews]

I have a working Zabbix setup where the Zabbix server is running correctly , and I’ve installed the Zabbix agent on two different machines using the same configuration steps. Both agents are enabled, running as verified by systemctl, and configured with the correct Server and ServerActive values pointing to the Zabbix server. The Hostname parameter in zabbix_agentd.conf is set to match the hostnames defined in the Zabbix UI. Despite this, only one of the agents is successfully performing active checks and appears green (Available) in the UI, while the other one stays gray and does not report active checks, even though it’s running fine in the backend.

[Markku]

What does the log file (zabbix_agentd.log) say?

Markku

[firdews]

Zabbix Agent Active Checks Issue – Log Analysis and Comparison

Hello,

Thank you for your response.

I’ve checked the zabbix_agentd.log file on the problematic agent machine, and I’m seeing repeated connection failures when trying to perform active checks:
2025-08-04 11:36:12 61413:20250803:011519.544 End of send_buffer():FAIL
2025-08-04 11:36:12 61413:20250803:011519.544 Unable to connect to [192.168.122.X]:10051 [cannot connect to [[192.168.122.X]:10051]: [4] Interrupted system call]


The agent appears to be running correctly in the background, but it fails to connect to the Zabbix server for active checks. However, I have another Zabbix agent — configured in exactly the same way — running on a different machine and successfully performing active checks against the same Zabbix server.

I’ve double-checked the following on both agents:

The Server and ServerActive parameters in zabbix_agentd.conf are correctly set to 192.168.122.X.
The Hostname matches the entry defined in the Zabbix UI.
The agent service is enabled and running via systemctl.
There are no apparent firewall issues or network blocks between the agent and server.
Port 10051 is reachable from both machines.
DNS resolution and routing are functioning normally.


Despite this, only one agent shows as available (green), and the other remains gray with no active checks received.

Could this be an issue with how the server handles multiple agents or host entries? I would appreciate any suggestions or if there’s anything else you recommend checking.

[firdews]

Zabbix Agent Active Checks Issue – Log Analysis and Comparison

Hello,

Thank you for your response.

I’ve checked the zabbix_agentd.log file on the problematic agent machine, and I’m seeing repeated connection failures when trying to perform active checks:
2025-08-04 11:36:12 61413:20250803:011519.544 End of send_buffer():FAIL
2025-08-04 11:36:12 61413:20250803:011519.544 Unable to connect to [192.168.122.X]:10051 [cannot connect to [[192.168.122.X]:10051]: [4] Interrupted system call]


The agent appears to be running correctly in the background, but it fails to connect to the Zabbix server for active checks. However, I have another Zabbix agent — configured in exactly the same way — running on a different machine and successfully performing active checks against the same Zabbix server.

I’ve double-checked the following on both agents:

• The Server and ServerActive parameters in zabbix_agentd.conf are correctly set to 192.168.122.X.
• The Hostname matches the entry defined in the Zabbix UI.
• The agent service is enabled and running via systemctl.
• There are no apparent firewall issues or network blocks between the agent and server.
• Port 10051 is reachable from both machines.
• DNS resolution and routing are functioning normally.

Despite this, only one agent shows as available (green), and the other remains gray with no active checks received.

Could this be an issue with how the server handles multiple agents or host entries? I would appreciate any suggestions or if there’s anything else you recommend checking.

[Markku]

2025-08-04 11:36:12 61413:20250803:011519.544 Unable to connect to [192.168.122.X]:10051 [cannot connect to [[192.168.122.X]:10051]: [4] Interrupted system call]

• Port 10051 is reachable from both machines.

What is the reason why you are saying that server port 10051 is reachable from the agent, when the agent logs that it cannot connect to it?

Markku

[firdews]

yesterday I was able to make active checks on another agent now in total I have two agents doing active checks. yet I added the third one the same way and the same configuration of the other two agents yet its not doing active checks

[Markku]

Do you have a network and/or firewall admin that could help you determining the reason why some of the agents cannot connect to the server?

Markku