graphing help please

[Mike]

On 2006-09-26, Buchan Milne <bgmilne at staff.telkomsa.net> wrote:
> --nextPart1939210.s1LOtCzRWx
> Content-Type: text/plain;
>   charset="iso-8859-6"
> Content-Transfer-Encoding: quoted-printable
> Content-Disposition: inline
>
> On Monday 25 September 2006 19:45, Mike wrote:
>> On 2006-09-25, Mike <mikee at mikee.ath.cx> wrote:
>> > I have hobbit setup and am working on a script to monitor all
>> > my platforms (redhat, solaris, irix, tru64, windows, etc). I
>> > have the script
>
> which script?

I am in a small development site that has one of many different platforms. I
have written a script in tcl that should/will run on all platforms to gather
the monitoring data. The tcl script is what I mention above. I'm happy to release
my tcl script once it is complete.

>> > running as a test on my local, windows=20
>> > workstation.
>
> Which client are you using ?

I am using the tcl script I mention above. I know the script is working as I
do get status messages from the tcl client into hobbitd on one of my linux
boxes.

>> > The status seems to be reporting properly from=20
>> > the script through the bb2.html page on to notifications. What
>> > I want to see next is a graph of these values. The current
>> > value is the simulated one minute load percentage from windows.
>> > I know this value is being sent to hobbitd since I'm getting
>> > notifications based on the colors (clear/green/yellow/red) I'm
>> > sending to hobbitd.
>> >
>> > Do I need to cycle hobbitd?
>> >
>> > Mike
>> >
>> > To unsubscribe from the hobbit list, send an e-mail to
>> > hobbit-unsubscribe at hswn.dk
>>
>> I have found the hobbit-graph.cfg file.
>
> You shouldn't need to touch anything here.

I agree, though I do not understand why the graphs are not showing up.

>> I'm trying to plot the=20
>> load percentage for the cpu. I have one line that's going in like:
>>
>> status HOST.cpu green
>>  12:00:50 up 7 days, 20:38,  0 users,  load average: 20 0 0
>>
>> Then another line that is:
>>
>> data HOST.cpu 20
>>
>> I'm wanting the 20 above, 20% of the cpu, to be plotted.
>
> I have only used the bbwin client for windows (bbwin.sf.net), not the Big=20
> Brother NT client. The bbwin client reports CPU usage to hobbit in the=20
> correct format for it to be graphed.

I like the bbwin on windows, but I want to run the same client everywhere,
hence the tcl script.

>
> Please provide some real details of what you are running on this client, if=
>=20
> you're not running bbwin. On all the unix hosts, the hobbit client should b=
> e=20
> used ...

I would like to use the hobbit client, but I do not want to compile the
client on every box. Compiling the client on every box is one of the few
weaknesses I find in hobbit. The hobbit clients do work on the three boxes
where they are used, graphs appear, etc. Status messages, events, etc., do
appear on hobbit from my tcl script, but graphs do not. I'm using a recent
hobbit, if not the latest. Hobbitd runs on a linux box (fedora core 5). The
other two linux boxes are of red hat flavor of different versions. When
in full production the client will report the status of irix, hp-ux, solaris,
windows xp, windows xp 64-bit, mac os x, dec, etc. I am happy to release my
script when it is complete.

Mike