performance monitoring

On May 8, 10:11*am, John Murtari <jmurt...@thebook.com> wrote:

> * * * * You may want to check out 'mrtg', it does all the
> graphing you want write out of the box. *It was first normally
> used for router traffic reports, but has been configured to
> plot any number and show historic graphs -- for more info
> check:http://oss.oetiker.ch/mrtg/

My recollection is that Mrtg was first, it was later considered
obsoleted by Cricket, and now they're both considered obsoleted by
Nagios. I still use Mrtg for quite a few forms of monitoring and it
meets all of my (admittedly simple) requirements.

I think the advantages of nagios and cricket lay in their
"enterprisey" features. Things like the ability to keep a deep
history, a sophisticated user interface, real back-end database,
multiple access levels to a web front end, configurable thresholds
with reporting and tracking when thresholds are exceeded, and so on.

DS

Douglas O'Neal schrieb:
> On 05/08/08 07:00, Björn Keil wrote:
>> David Schwartz schrieb:
>>> On May 7, 9:45 pm, Peter <one2001...@yahoo.com> wrote:
>>>> hello,
>>>> is there a graphical tool that can show the usage of CPU, memory for
>>>> the
>>>> past week? most tools I used are command line utilities. not sure if
>>>> there is a graphical tool that I can see online. or is there a script
>>>> that I can write the performance data to a file and then display in GUI
>>>> in another linux box?
>>>> thanks.
>>>> '
>>>
>>> Nagios is probably the most commonly recommended free tool for doing
>>> this. It can monitor anything that can be expressed as a number -- CPU
>>> usage, free disk space, network traffic, number of processes running,
>>> page faults, whatever.
>>>
>>> DS
>>
>> Nagios is for monitoring the current state, it doesn't provide history
>> curves.
>>
>> Personally I use cacti for graphs and stuff, but that might be a
>> little too heavy if you just wanna monitor a single one host. It
>> requires you to have a webserver and a database on one host (though
>> not necessarily the same) and SNMP installed on the machines you wanna
>> supervise.
>> http://www.cacti.net/

> It is trivial to have nagios log performance data to a flat file or a
> database. That data can then be fed to your graphing tool of choice.

That may be true but it's still easier to use a program for a given
purpose that were made for THAT purpose that just a purpose that shares
some similarities. ;)

On 05/09/08 06:51, Björn Keil wrote:
> Douglas O'Neal schrieb:
>> On 05/08/08 07:00, Björn Keil wrote:
>>> David Schwartz schrieb:
>>>> On May 7, 9:45 pm, Peter <one2001...@yahoo.com> wrote:
>>>>> hello,
>>>>> is there a graphical tool that can show the usage of CPU, memory
>>>>> for the
>>>>> past week? most tools I used are command line utilities. not sure if
>>>>> there is a graphical tool that I can see online. or is there a script
>>>>> that I can write the performance data to a file and then display in
>>>>> GUI
>>>>> in another linux box?
>>>>> thanks.
>>>>> '
>>>>
>>>> Nagios is probably the most commonly recommended free tool for doing
>>>> this. It can monitor anything that can be expressed as a number -- CPU
>>>> usage, free disk space, network traffic, number of processes running,
>>>> page faults, whatever.
>>>>
>>>> DS
>>>
>>> Nagios is for monitoring the current state, it doesn't provide
>>> history curves.
>>>
>>> Personally I use cacti for graphs and stuff, but that might be a
>>> little too heavy if you just wanna monitor a single one host. It
>>> requires you to have a webserver and a database on one host (though
>>> not necessarily the same) and SNMP installed on the machines you
>>> wanna supervise.
>>> http://www.cacti.net/
>
>> It is trivial to have nagios log performance data to a flat file or a
>> database. That data can then be fed to your graphing tool of choice.
>
> That may be true but it's still easier to use a program for a given
> purpose that were made for THAT purpose that just a purpose that shares
> some similarities. ;)

I use a data-gathering tool and add to it a graphing capability. You
use a graphing tool and add in the data gathering. I'm not sure that
I see a huge difference.

On Fri, 09 May 2008 08:16:30 -0400, Douglas O'Neal wrote:

> On 05/09/08 06:51, Björn Keil wrote:
>> Douglas O'Neal schrieb:
>>> On 05/08/08 07:00, Björn Keil wrote:
>>>> David Schwartz schrieb:
>>>>> On May 7, 9:45 pm, Peter <one2001...@yahoo.com> wrote:
>>>>>> hello,
>>>>>> is there a graphical tool that can show the usage of CPU, memory
>>>>>> for the
>>>>>> past week? most tools I used are command line utilities. not sure
>>>>>> if there is a graphical tool that I can see online. or is there a
>>>>>> script that I can write the performance data to a file and then
>>>>>> display in GUI
>>>>>> in another linux box?
>>>>>> thanks.
>>>>>> '
>>>>>
>>>>> Nagios is probably the most commonly recommended free tool for doing
>>>>> this. It can monitor anything that can be expressed as a number --
>>>>> CPU usage, free disk space, network traffic, number of processes
>>>>> running, page faults, whatever.
>>>>>
>>>>> DS
>>>>
>>>> Nagios is for monitoring the current state, it doesn't provide
>>>> history curves.
>>>>
>>>> Personally I use cacti for graphs and stuff, but that might be a
>>>> little too heavy if you just wanna monitor a single one host. It
>>>> requires you to have a webserver and a database on one host (though
>>>> not necessarily the same) and SNMP installed on the machines you
>>>> wanna supervise.
>>>> http://www.cacti.net/
>>
>>> It is trivial to have nagios log performance data to a flat file or a
>>> database. That data can then be fed to your graphing tool of choice.
>>
>> That may be true but it's still easier to use a program for a given
>> purpose that were made for THAT purpose that just a purpose that shares
>> some similarities. ;)
>
> I use a data-gathering tool and add to it a graphing capability. You
> use a graphing tool and add in the data gathering. I'm not sure that I
> see a huge difference.

I've heard good things about zenoss. It's supposed to be a sort of
combination of what nagios does and what cacti does - plus a little more,
because with zenoss you can, using the same tool, see what the load
factor was right before a crash - for example.

I've set up nagios twice and cacti once, but next time around I'll likely
give zenoss a try.