agent.ping is checked once, then goes silent

[ThomasGU]

Hi,

I'm having an issue with my Zabbix 4.2.1 installation on RHEL 7.2 with PostgreSQL 9.2.13 :

All host monitored with the Zabbix agent trigger the "Unreachable for 5 min" alert, yet the zabbix_get -k agent.ping command returns 1 on all of them and the refresh time is lower than 5 min.

Latest data indicates "Up (1)", but the last check happened when the server restarted. So when I restart the zabbix-server service, the alerts are RESOLVED for 5 mins then go back to PROBLEM state for lack of data during 5 min :

(Just after restarting the service at ~12:10)

(After 5 min)



Using the "Check now" feature for the Agent Ping item doesn't generate a new value on any of the hosts and the server monitors successfully all other agent items :


With Zabbix server DebugLevel on 5, I seem to get successful polls :

In get_value_agent() host:'hostname' addr:'address' key:'agent.ping' conn:'unencrypted'
Sending [agent.ping]
get value from agent result: '1'
End of get_value_agent():SUCCEED
End of get_value():SUCCEED

This looks like a bug but I'm probably missing something so any help is greatly appreciated, thanks in advance


Thomas

[ThomasGU]

Solved, the "Discard unchanged" option was enabled on the Agent app template....