agent.ping against Zabbix agent (active) not returning anything

[TheCookieMonster]

I have set up an Item on a host with the following config:
Type: Zabbix agent (active)
Key: agent.ping
Type of information: Numeric (unsigned)
Data Type: Decimal
Units:
Update Interval: 5 secs
History: 180 days
Trend storage: 365 days
Store value: As is
Show value: As is
Applications: None

However when I look in latest data graph or latest data values for this item, there is no data, despite the agent being active and providing values for other items.

Can anyone help?

[jusavard]

Hi,
Can you run zabbix_get with the key like this :

[zabbix@zabbix-server ~]# zabbix_get -k agent.ping -s TARGET_HOST

Also, are you sure you want a graph with the value of agent.ping ? This key only return 1 ( if agent ping) or 0 ( if not). Could "icmppingsec", which return the time of each ping, be what you'd be looking for ?

[TheCookieMonster]

Hi,
Can you run zabbix_get with the key like this :

[zabbix@zabbix-server ~]# zabbix_get -k agent.ping -s TARGET_HOST

Also, are you sure you want a graph with the value of agent.ping ? This key only return 1 ( if agent ping) or 0 ( if not). Could "icmppingsec", which return the time of each ping, be what you'd be looking for ?

Unfortunately I dont have access to the Zabbix server to run that, only have access to the GUI.
And Yes, I essentially want to know if the server is up or down, hence the 1 or 0 check

[anewark]

I have set up an Item on a host with the following config:
Type: Zabbix agent (active)
Key: agent.ping
Type of information: Numeric (unsigned)
Data Type: Decimal
Units:
Update Interval: 5 secs
History: 180 days
Trend storage: 365 days
Store value: As is
Show value: As is
Applications: None

However when I look in latest data graph or latest data values for this item, there is no data, despite the agent being active and providing values for other items.

Can anyone help?

Is this on a linux host? SELinux and Firewall block agent ping frequently. I'd suggest looking at your exceptions for SELinux and the firewall if that's the case, or maybe turn them off and see if it resolves the issue. If it's a Windows host make sure the firewall has the right exceptions in place too.

[jusavard]

Unfortunately I dont have access to the Zabbix server to run that, only have access to the GUI.
And Yes, I essentially want to know if the server is up or down, hence the 1 or 0 check

ASIK, agent.ping will monitor if the zabbix agent is up and running. Which doesn't mean you server is up/down hence the icmppingsec. Nevertheless, the "agent.ping no data thing" intrigue me...

Can you double check if the item is "enabled" ?

Also, when I look on my my side for this item I see :

Type: Zabbix agent
not
Type: Zabbix agent (active)

I thinkthere is a diffrence between these 2. If i remember right, the "non-active" agent mean the server pull the info from the client and "active" agent mean the client push the info to the server.

[allexpetrov]

TheCookieMonster , If you want to ping the host, just try the Simple Checks ( https://www.zabbix.com/documentation.../simple_checks ) type with icmpping[] key, which will return 0 or 1. Then choose show value as Service State and you will receive the value in Human readable format.

[kloczek]

FYI agent.ping[] item is for monitor passive agent.
In case of use active agent setup agent process does not listen on tcp port because in such setup agent initiates communication to prx/srv.

Typical approach to monitor active agent activity is choosing one of the item with lowest sampling rates and use nodata() trigger function.

https://www.zabbix.com/documentation/3.2/manual/appendix/triggers/functions

Setting ServerActive and Server in agent setting allows you to use agent.ping[] item but it it forces to open srv/prx->agent communication but such setup is in contradiction if you want to keep network layer data flow settings secure.

What it means "secure" in this context?
If lets say someone will be able penetrate single host monitored over zabbix and if prx/srv process will have some 0-day security bug penetration of the single monitored node will allow to penetrate srv/prx as well and because passive agent setup must have opened communication from srv/prx this could be used penetrate rest of the monitored infrastructure.