Failover on host agent when performing Ceph supervision across multiple Proxmox PVE
Hello,
Recently, I configured the Proxmox template on my Zabbix to monitor my PVE servers. It works perfectly fine, but I do have one concern about it. I could be unaware, but as far as I know, failover for Zabbix agent is not a thing. This is quite bothersome when you need to supervise multiple PVE. Let me give you an example:
Let's say you have two PVEs:
PVE-A and PVE-B
You configure your host on Zabbix to use PVE-A (the default) and add PVE-B to the same host, meaning every check will use PVE-A as the main host/interface. Everything works fine—great!
Now, imagine PVE-A goes down. Since it is down, your checks no longer work. Meanwhile, on the Proxmox side, PVE-B has taken over, becoming the active Ceph-MGR, ready to respond to Zabbix-Agente2 requests (REST API).
And here is my problem... I want Zabbix to automatically switch between the PVE-A agent and PVE-B agent for the checks/host when such a situation occurs, directly alerting me that one of my PVEs has gone down.
Thank you very much for your help.
Recently, I configured the Proxmox template on my Zabbix to monitor my PVE servers. It works perfectly fine, but I do have one concern about it. I could be unaware, but as far as I know, failover for Zabbix agent is not a thing. This is quite bothersome when you need to supervise multiple PVE. Let me give you an example:
Let's say you have two PVEs:
PVE-A and PVE-B
You configure your host on Zabbix to use PVE-A (the default) and add PVE-B to the same host, meaning every check will use PVE-A as the main host/interface. Everything works fine—great!
Now, imagine PVE-A goes down. Since it is down, your checks no longer work. Meanwhile, on the Proxmox side, PVE-B has taken over, becoming the active Ceph-MGR, ready to respond to Zabbix-Agente2 requests (REST API).
And here is my problem... I want Zabbix to automatically switch between the PVE-A agent and PVE-B agent for the checks/host when such a situation occurs, directly alerting me that one of my PVEs has gone down.
Thank you very much for your help.