Hello,
First, I’m sorry for the very generic title of this new topic. I didn’t really find a good one for my question.
Here is my situation:
I’m using physical servers from a Cloud provider. This Cloud provider push public metrics from these servers, like the processor/motherboard/room… temperatures
On my Icinga2 configuration, each physical server is declared as a Host object. Pretty basic
On each host, I’ve created a special service that checks these metrics. It works perfectly and I’m getting alert when the temperature is too hot.
But…
Sometimes, the API I call to retrieve the metrics is down. As I’m not the maintainer of this API, I can’t do anything. So, when it happens, I receive a ton of notification because all of the checks are unknown.
My idea:
I’d create one host, that is the metrics API itself. I’d apply a simple check_http on it.
On this host, automatically create one service for each Host object (with the flag “physicalServer” = true).
Each service will check the temperature of one physical server.
By doing that, if the metrics API is down, I’ll receive only one notification, not a ton from all of the check temperature service.
Is that possible?
Thanks for reading,
Kevin