[AusNOG] Network Management and Tools

[Tony]

--- On Mon, 5/7/10, David Hughes <David at Hughes.com.au> wrote:
> 
> This is one thing that amazes me about FOSS monitoring
> systems.  Everyone thinks its normal and totally
> acceptable to have 2 different systems for realtime
> monitoring and historical reporting.  
>
> I know we all do it but why do we all think it's perfectly
> fine to collect almost every single monitoring metric twice
> ?  What an amazingly inefficient way of doing
> this.  Problem is I can't find anything that I'd be
> happy to use as a replacement for nagios and cacti. 
>

I know it's not FOSS in any way, but has anyone considered PRTG ? (yes I know it's closed source, only runs on Windows and costs real $) 

Without being a walking advert for it, PRTG supports recording both historical data and alerts/notifications on current polling results (you can also alert on trends, you choose the interval). Something else that we like about it is that you can monitor at intervals that aren't restricted by how often CRON can run a task (yes, I know there are workarounds for this). Meaning that if we want to we can monitor things down to 1 second intervals. All of the data is recorded in a (proprietary) DB format at full resolution so I can produce a graph from 0300-0500 from 5 weeks ago at 5 second resolution if I wish to. Yes, this takes up disk space proportional to how many items you have and how long you keep the data for, but how cheap is disk space these days anyway ?

We use it and are very happy with it and as such if anyone has questions I'm happy to try and answer them. I recently (about 18 months ago) tried to find a FOSS alternative that wouldn't involve me doing a LOT of hacking and customisation myself and couldn't find anything so we're still using PRTG.


Going a bit off-topic (although I think the OP did ask about troubleshooting utilities) I use a small utility I picked up simply called "stg.exe" (SNMP Traffic Grapher) for diagnostics sometimes. This is a small Windows (yes I know) executable that will just plot two OID's in realtime on your screen. What I really like about this utility is that I can set the polling interval as low as I need to diagnose network issues. I have set a 100ms (yes, 10 times per second) poll on it to diagnose problems on an ATM interface. The readme that comes with it says it is free, but not open source. It's basic, but is quick to use and does the job I need it to.



regards,
Tony.

(yes, flame on about how PRTG isn't free, doesn't run on a free OS, etc. I think the cost of it is fairly small compared to some of the other paid products out there. Having a Windows PC/server to run it on isn't such a big issue is it ?)