Re: [tlug] Server monitoring/graphing

[Attila Kinali <attila@example.com>]

On Mon, 10 Oct 2016 12:04:18 +0900
Kalin KOZHUHAROV <me.kalin@example.com> wrote:

> On Mon, Oct 10, 2016 at 8:46 AM, Tim Osburn <tim@example.com> wrote:
> > On Mon, 10 Oct 2016, Attila Kinali wrote:
> >
> >> I recently set up an NTP server and wanted to add that into my "monitoring"
> >> and of course I hit the limitations of mrtg.
> >
> > Curious which limitation of MRTG you hit?

The biggest problem is that MRTG is unable to handle negative numbers.
Then there is the issue that MRTG is not made to plot more than 4 values
in the same graph. 


> +1 Yep, without much input, only general recommendations are possible.

Well, I am asking a very general question :-)
I wanted to see, whether there is a modern replacement for MRTG.
E.g. I wouldn't mind to get a frontend like smokeping has, where
one can easily zoom in.


> Have you tried Cacti or Zabbix? This is what some of my customers use.

I was under the impression that most people moved away from Cacti again.
I haven't had a look at Zabbix yet.

> For bigger shops monitoring is often part of log management, so
> performance data goes in the syslog stream and is picked up by
> Splunk/ELK/whatever.

Yes, that's my problem. I don't need to and don't want to go that big.
I'd like to keep it simple and easy to configure. I am not a professional
in this area, my needs are relative easy to fullfill. I am not worried
about managing thousands of machines either. Heck, none of my MRTG setup
monitors more than 3 machines (and this one by accident only).

Thanks for the answers. I will have a closer look at Zabbix and Monitorix
and come back with more questions :-)


			Attila Kinali

-- 
Malek's Law:
        Any simple idea will be worded in the most complicated way.