Monitoring hosts and suppressing child host alerts/triggers

[isparkles]

Having some trouble setting up alerting based on site devices or suppressing alerting based on some host being unavailable.

Our specific example is we have many sites, all the sites have a UPS and router, and a bunch of switches and servers. If the UPS or the router is unavailable, all the sites devices right now start alerting that they are unavailable, which is understandable.

In another product you could link host groups and suppress alerts of a child group if members of the host group are unavailable.

How would we do this in zabbix? Or is there another way of doing something like the above?

Many thanks

[tim.mooney]

Triggers in Zabbix can depend upon other triggers, take a look at the documentation for triggers and dependencies.

In my environment, I set up network topology dependencies for the paths from Zabbix to each of our networks within each of our datacenters, and then I have the base host monitor (Zabbix agent ping, in the case of systems that run the agent) depend upon the appropriate network gateway.

I recommend you monitor your network topology dependencies at a higher frequency than you monitor your endpoints. This reduces (but does not eliminate) the chance of Zabbix thinking an endpoint is down before it has run the check that actually determines that it's not an endpoint that's down, it's some part of the topology that's unreachable.

Unrelated to topology dependencies but since you mentioned multiple sites, you may also want to consider proxies at some of the sites, if you are bandwidth-constrained to any of them.

[gofree]

simillar topic

https://www.zabbix.com/forum/zabbix-help/435580-event-correlation