Distributed monitoring

[Dre]

I'm looking to build a new Zabbix platform, and I have a question about distributed monitoring.

One of the things I am looking to do, is availability monitoring. This would mainly rely on ping, but in some cases also other protocols. And I'm looking to do this in a distributed manner, in such a way that the checks get performed from all monitoring locations. And an alert should be triggered not only when all locations report "down", but rather when any location reports "down".

Is this possible - and can anyone point me at some resources on how to set this up?

The main reason I'm asking this, is because in the documentation I'm mainly reading about things like redundancy and scalability (e.g. adding multiple monitoring servers for more capacity), but I don't really see anything about the functionality that I just described.

Just to be clear: the purpose of such a setup would be to make sure that a host is globally reachable - since monitoring from a single location would only confirm that the host can be reached by that monitoring server. So what I'm looking to do is to confirm the availability over as many different routes as possible, rather than from only a single route.

Thank you.

[BDiE8VNy]

How about putting a Zabbix proxy on each monitored location and creating then in Zabbix one host per location?
The checks (Items and Triggers) may be located in a single Template that then could get assigned to each of these hosts.
Variable information, like domain names or IP addresses, should be given by macros.

[Dre]

Thank you for the suggestions. That certainly sounds like the right direction.

The first thing that comes to mind though: say we'd have 10 different monitoring locations (e.g. we monitor all services from 10 different locations). Then wouldn't this approach also generate 10 alerts when a service is completely unreachable? Is there a way to prevent that from happening?

[BDiE8VNy]

Zabbix triggers are not related to hosts but related to items. So you could additionally add a single trigger considering several items coming from different hosts.

To avoid having all the individual triggers active while the "complete unreachable" trigger is true, one could make the individual ones 'depending' on the 'complete' one. (See Trigger dependencies)

Instead of adding a dependency for each individual trigger, one can add this dependency on template level (when individual trigger were inherited from a template)
The dependency gets then inherited as well.

[Dre]

Ahh that makes a lot of sense. I wasn't aware that Zabbix has such features - that's awesome. I've been working with many different monitoring systems over the past 20 years, but either the systems were very simplistic, or the systems were advanced and I never made it beyond the very basic configuration.

That leads me to a related question: setting up a reasonably advanced/decent monitoring system has been on my to-do list for literally years, and it's the kind of thing I'll probably never get around to put some time into (at least not to really do it properly). With that said: can you tell me what would be an appropriate place to ask people for paid services (e.g. someone who would take care of installation/configuration and perhaps some custom development to monitor some specific things)? I know that there are various sites where you can hire freelancers - but I'm wondering whether there's perhaps a place in this community where I can ask?

[BDiE8VNy]

Well, Zabbix SIA offers the full range of services around Zabbix.
See Zabbix SIA Services Overview

Support can also be obtained from various certified partners around the world.
See Zabbix Partners