jc,
I was with you right up until you said "the raw data isn't transmitted up."
Does this mean that even if I were to populate the clientlog on the client
(say, via the script you referred to), that data would not be transmitted
back to the server?
Thanks,
David
On Mon, Nov 16, 2015 at 1:45 PM, Japheth Cleaver <user-87556346d4af@xymon.invalid>
wrote:
On 11/16/2015 10:29 AM, David Welker wrote:
I am interested in using this to create a testbed. From what I can tell,
it only sends a message that contains the [date], [uptime], and [free]
sections of the clientlog to the server, so how are the rest of the columns
created, by the xymond daemon?
Is there any simple way to add ones own columns to this structure and have
them show up on the server?
Thanks,
David
Hi David,
In a normal, server-configured, installation, most of the columns
representing client-pushed data like that are generated by the
xymond_client daemon running on the server. It receives the clientlog and
generates status messages which show up as columns.
You can add sections in, but something needs to be "listening" to know how
to interpret that data, decide what color the resulting column is, and send
the status message in to xymon.
In client-configured ("local config") mode, the same thing is basically
happening, except that xymond_client is running on the client machines
themselves, and the raw data isn't transmitted up.
To add raw data to the clientlog, you can edit the xymonclient-${OS}.sh
script yourself, or (depending on whether it's packaged or from source),
create a ~/local/ directory containing small executables that print out
whatever data you'd like to include.
Alternatively, if you can just create the status message yourself on the
client too. Sending details via the client message isn't a strict
requirement.
HTH,
-jc