NMS vs Grafana

[Eric Faden]

So continuing on my path of improving my network monitoring...

I have a network I am trying to setup a new monitoring stack on. Right now my plan is...

• Syslogs -> FluentD -> ElasticSearch -> Graylog
• SNMP/NetData/fping -> Telegraf -> Prometheus -> Grafana
• IPFix -> vFlow -> Telegraf -> Prometheus -> Grafana

The issue I am having is that I still can't figure out if that setup will cover the last bit I need. I want to have NMS (Zabbix/OpenNMS/LibreNMS) like functionality in the above stack. I just need something that will check for connectivity, "up-ness" of the routers/switches/etc. Things that are more boolean status vs time based data. I want a dashboard similar to a NMS just showing that type of data so I can see at a glance if anything in the network itself is broken.

Does that make any sense? Can I do this in the above setup? Thus far I know I could do

LibreNMS -> Telegraf (via Graphite Protocol) -> Prometheus -> Grafana

Or

OpenNMS -> Grafana (directly as a data source)

or

Nagios Checks -> Telegraf (via exec plugin w/ Nagios Data Type) -> Prometheus...

or even

Zabbix -> Grafana

Thoughts?

-Eric

 

[Eric Faden]

Basically the goal is to replicate a NMS style boolean check into the telegraf stack... so I don't ALSO have to have a NMS running.... but open to any thoughts?