Availailbility status
I use 1.1Alpha7 and I have problem with availailbility status. All devices and hosts show status monitoring. That's OK. But some of them show availability status unknown.For example all snmp devices show status unknown and some hosts running zabbix agentd not all show also unknown. I also add item status with simple check to snmp devices and it doesn't return value.Also tried with icmpping key - simple check which returns value but availailbility status also doesn't change I am out of ideas how to solve this problem.
-
-
Ditto
I too am having this issue. I have many routers that I monitor (not Ciscos with SNMP, but ones I do ping tests for). I have "Availability: Not available" and "Error: Cannot connect to [rtr-1] [Connection refused]" in the Host configurator. I am currently simply ignoring these messages, but the screen has ~20 red error messages on it and it bugs me when I look at it. Maybe besides "Monitored, Not Monitored, Template, a "Clientless" option or maybe "Passive" (although that would lead to the question, is the client passive or the server? So maybe "Clientless" would be a nice way to put it? Maybe put "Clientless" in Yellow instead of green then too? Just a thought.
Back to my original concern, is it ok to simply ignore these errors? Or is there a "proper" way to set up clientless systems so I don't get these errors??
Thanks.
By the way, if you read this Alexei - Thank you for all your hard work, Zabbix is FAR better than Nagios!
-
It all depends on how you define availability. It seems that currently it's defined as "Ability to connect to a ZABBIX or SNMP agent". So, if ZABBIX server cannot connect to the host it sets availability trigger to TRUE.
However, you may define availability as status of ICMP ping. So, if ZABBIX agent can ping a host, the host will be available. Otherwise - not available.
Do not rely on default trigger expressions! Define your own availability criteria.Comment
-
perhaps for each host, it would be fine to define an item or a trigger which will be taken as the availability status
for example, if a host have only an icmpping item, the status will be unknown.
but if in the config of this host one can define that the status of this host is linked to the item icmpping (a value is returned, status is OK, no value, status NOK) perhaps it would solve problems.
but as alexei said, this status is only in the config pages, and it depend on what it means. only admins, who are supposed to know what it means (if the documentation were as complete as the subway this morning
), are allowed to see it. and as alexei said again, make your own availability triggers.
Your triggers should be your references to determine the status of a host.
and YES, zabbix has a much greater potential than nagios. hope it will grow.Comment
Comment