[TriLUG] Easy SNMP monitoring of Linux servers

[Kevin Otte via TriLUG]

I'm fond of collectd for statistics gathering. It defaults to writing to
RRD files locally for which you'd need to have something to present the
graphs. Having it write to a graphite instance and then using grafana to
display graphs is much more visually pleasing. Also, grafana recently
implemented alerting.

carbon and graphite can be a bit tricky to get going, but in my opinion
it's still much easier than fighting with cacti. The main thing to keep
in mind is the metrics are pushed back to the collection point rather
than polled. This will likely affect your security stance.

-- Kevin

On 05/26/2017 01:18 PM, Ron Kelley via TriLUG wrote:
> Greetings all,
> 
> I am looking for a easy-to-use and easy-to-install tool to monitor system performance for Linux.  I just went through the cacti installer again and remembered why I dislike the tool so much.
> 
> I was looking at Zabbix but was wondering if there is something else to try.  My main requirements are RAM, CPU, HDD space monitoring and graphing of Ubuntu servers.  Keeping historical data is very important - need the ability to look back over a few months of data.  I don’t care if the tool requires SNMP or an agent.  Nor do I need a tool with 1,000 features.  Just something dead simple to get performance stats from servers.
> 
> Thanks.
>