Zabbix, failover monitoring?

Question

Zabbix

Greetings!
Prompt as it is possible to implement monitoring of transition of branches to a reserve channel.
Those. the branch sits on its main channel, does not know the trouble, it has IP 192.168.1.20 (l2vpn from providers for all branches). And then the channel suddenly dropped, the branch goes to the reserve (usb 3g modem) and connects a regular vpn and gets the address 192.168.10.30 (when connected via VPN, it will always have this address). Everything is back to normal, everything works. But zabbix continues to consider this branch as inaccessible and not having access to our network, which is quite natural, because. accessibility is monitored by the main IP.
And now I just can’t figure out how to teach zabbix to see this transition to the backup channel.
Of course, you can create two network nodes for each branch, but then there will be a lot of notifications about the unavailability of network nodes, which is very inconvenient.

Answer 1

Zabbix since version 2.0 supports assigning multiple interfaces to a single host. However, I'm not sure it will help you much. You need to switch from channel checks to active agent checks via the same agent.ping.

Answer 2

use UserParameter. for example, UserParameter=link.down[*],(ip link | grep $1 | grep DOWN > /dev/null) && echo 1 || echo 0where $1 is the name of the network interface

Answer 3

As an option, I can offer to raise a vpn tunnel between zabbix and the server, then the server will always be available at a private address.
I have a large geographically dispersed vpn network going to the same server. So I launched my own zabbix for monitoring remote hosts in this way. Then it doesn't matter how all clients get to the vpn server, because they are always available inside the tunnels.

Answer 4

I would try to monitor not the current IP addresses, but the logs of the VPN server.
IMHO, it is not important for the authorities to monitor failures in real time, but the final report for a certain period.
The logs will allow you to understand both the number of failures and their duration.

Answer 5

I used a script that pinged both routes, in the event of a link failure, the route was rebuilt to a backup one and a letter was written to the admins.