ping check needs interface - the given interfaces wont work

[[email protected]]

Hi,

I am pretty new to Zabbix

I started a Zabbix Server and Created a Host to monitor

for the first step I thought, I do what I did for the first steps in all the other monitoring tools, like nagios. and set up an easy ping check.
What was very easy, and done in some minutes with the all the others, but in Zabbix that seems to be an impossible task..

in a lot of places and in this forum, its always said :
put under interface anything, it only needs the Host or IP.
So I did. (Interface type "Agent" with the Hosts IP and the default port 10050 I didnt change)
Now I have in the list of hosts a host where in the column availibility is a grey ZBX Icon, which seems to indicate, that ZBX its not available, which is not surprising.
and in the column interfaces a host IP and a port (a port which does not exist and I wont use for this host).. it really feels : this cannot be ... this is confusing and something must be wrong.

And also it was not working. The error message : "Ping item must have target or host interface specified" is still there and I have a failed host.




as an Interface I can only choose from a list: JMX, AGENT, IPMI, SNMP. None of them IMHO are actually proper if it only comes to an agentless ping check.
I do not understand what interface is necessary for a ping check, and why I can only choose such with a complete different purpose - I wasnot lazy and I tried all of them: the error message "ping must have target.." was in all cases still present.


Even I found something like this (which is sadly much more confusing one has to do so for a ping check in monitoring system)

https://www.zabbix.com/forum/zabbix-patches/413761-displaying-the-icmp-status-of-a-hosts

What do I do wrong? What do I not understand ? Which button, select box, entry box, checkbox I obviously fail to see.

BR,
Andrew

[cyber]

Screenshot of your item config would be helpful...

[[email protected]]

Hello,

yes, I know, I am the one, who usually asks for a screenshot.

in this particular situation its better we close that post. I could make it work. And I am way further in my configurations.. meanwhile the linux and windows server got agents, and I made myself a nice overview dashboard derived from the default and even built up a little map, like I have seen in some KnowHow Videos, one can find in a special forum section here.
Next step is snmp configurations with the already discovered (via discovery service) network components (for which I have configured for me already at least icmp checks)
Its a pitty, that ICMP Ping is not considered as "Available Host", and is shown under "unknow". All the new IoT Devices in our new business building and some dumb network components I will always have shown as unknown. very strange, since IoT can really easily outnumber the server components nowadays. In my opinion zabbix should consider here a change of philosophy and strategy. In a parallel work I will test out a prometheus/grafana combo. I already know very well grafana with influxdb as my visualizer for the collected data of my private home assistant installation..


For readers, who stranded in my post: I can confirm, that a ping check (ICMP Template) is working if e.g. in the host config as interface the agent interface is configured (no matter what port, I use 0 for IoT, and 10050 for servers, which will get a zabbix agent soon)
the "item" in the host must be configured for this host interface.
Usually this is done automatically, if you configure it straight forward. If you play too much with the configurations, like me as newbie, than the entry host interface could be empty in the item config.
if the item host interface enry is configured to the agent interface, and the agent interface is configured in the host configuration, then all will be fine and the item can be enabled and works

[cyber]

Those "availability" icons show availability specific type of interface... agent, snmp, jmx etc... ping could be used with any of those... So I kind of understand, why they require interface. By default you ping that host... But then again, if you use it as simple check and with parameters (target), its performed by proxy anyway and towards the target and having a host interface is kind of irrelevant...
Dev ideas can always be submitted to https://support.zabbix.com/browse/ZBXNEXT or check, if someone has already proposed something similar and vote it up ..

[Hasabadusa23]

Didn't work for me, but what worked was this here

https://www.tarball.ca/posts/zabbix-...ace-specified/

• Go to Data Collection > Templates
• Locate the "ICMP Ping" template and click on Items
• There are three items: ICMP response time, ICMP ping and ICMP loss - append the following to the end of each key: [{HOST.HOST}]

The keys should look like this:
icmppingsec[{HOST.HOST}]
icmpping[{HOST.HOST}]
icmppingloss[{HOST.HOST}]

[willzb]

Didn't work for me, but what worked was this here

https://www.tarball.ca/posts/zabbix-...ace-specified/

• Go to Data Collection > Templates
• Locate the "ICMP Ping" template and click on Items
• There are three items: ICMP response time, ICMP ping and ICMP loss - append the following to the end of each key: [{HOST.HOST}]

The keys should look like this:
icmppingsec[{HOST.HOST}]
icmpping[{HOST.HOST}]
icmppingloss[{HOST.HOST}]

It work in zabbix v7.0.0