lun monitoring

[Tom Ammon]

Russell Adams wrote:
> On Mon, Feb 02, 2009 at 07:10:45PM -0700, Tom Ammon wrote:
>   
>> Russell,
>>
>> Cacti is pretty SNMP-centric, but in our environment that is about all 
>> we are using it for anyway. I'm no cacti expert, but to me, that's the 
>> beauty of it - I don't really know the inner workings of cacti, and I am 
>> not a programmer or scripter, but I got it up and running pretty quickly.
>>     
>
> SNMP is a great place to start, and very open. Its certainly more
> reliable than the CIMOM implementations I see.
>
>   
>> I'm not sure if you would call it autodiscovery, but cacti does do an 
>> snmpwalk on the devices that you specify, and the pre-built data 
>> collection methods that come with it are designed for getting snmp 
>> interface statistics. You can, of course, add other data collection 
>> methods, but out of the box, it is basically an interface traffic 
>> grapher. You still have to manually input each device that you will 
>> collect data for. Once you have specified the basic host information, it 
>> gives you a table showing all of the interfaces on that device and a 
>> checkbox for each item that can be graphed.
>>     
>
> Torrus is configured by feeding it a list of IP addresses and it
> identifies the device and sets up all the counters to be
> monitored. The detail is very good, more than just interface stats.
>
>   
>> To be fair, though (and this applies to nagios as well as cacti) most of 
>> the effort you put in to setting up a monitoring solution is a one-time 
>> thing. It takes time to input all of the devices, but for the most part 
>> once the devices to be monitored are specified, that work is over. I 
>> think people incorrectly place a lot of emphasis on this or that 
>> product's autodiscovery function. Cacti's interface makes it really easy 
>> to maintain the configuration, and I think that is a bigger win than 
>> autodiscovery.
>>     
>
> I consider autodiscovery to be absolutely critical. Maintaining a
> handfull of machines is one thing, hundreds or thousands or machines
> outside of your control are another. I wrote NACE to allow me to
> perform fast autodiscovery for Nagios, and I've been pleased to couple
> it with Torrus so they both have the same list of hosts.
>   
That is probably where our differing environments cause us to need 
different things. In my environment I monitor hundreds, but not 
thousands of devices. And they are all in my control. If I worked for a 
large ISP, I'm sure I would see things differently.

With Torrus, on a router, for example, what kind of detail would it 
typically give you outside of the normal interface statistics? Would it 
be able to discern cpu usage, memory usage, etc. without you specifying 
some kind of template for it to use as a reference?

Cacti has sort of solved this with their data templates. For example, 
there is a Unix Host Template that you can download and then apply to a 
device, and it gives you all of the parameters that are built in that 
template, for example, cpu/mem/disk. But the author of the template had 
to know the OIDs (and use the correct OIDs). It wasn't really 
autodiscovered.

Tom


-- 
-------------------------------------
Tom Ammon
Network Engineer
Mobile: 801.674.9273

Business Card at http://tomsbox.net/bizcard_TomAmmon.jpg

Center for High Performance Computing
University of Utah
http://www.chpc.utah.edu


------------------------------------------------------------------------------
Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
software. With Adobe AIR, Ajax developers can use existing skills and code to
build responsive, highly engaging applications that combine the power of local
resources and data with the reach of the web. Download the Adobe AIR SDK and
Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null