Currently, I’m with a rather small company with one site. So, my page
layout is pretty simple. There are pages for production, development,
network, storage, and external(DMZ). In production systems Unix boxes are
all together, we have few compared to Windows, and Windows boxes are
separated by function, i.e. database, file, web, email. Other pages are
much the same way. The only reason I separate Unix from Windows, though, is
because I think that putting the Unix stuff, which uses certain columns, and
the Windows stuff, which bbwin gives certain columns is ugly because of the
empty test columns. I use group-compress.
Previously, I was with a rather large employer. At that company, there
were pages for sites, and then subpages like described above.
Jamison Maxwell
user-87d336c3dce6@xymon.invalid
From: xymon-bounces at xymon.com [mailto:xymon-bounces at xymon.com] On Behalf
Of Steve Holmes
Sent: Wednesday, April 04, 2012 2:51 PM
To: Ralph Mitchell
Cc: Xymon Email List
Subject: Re: [Xymon] Thought Process for Xymon Page Layout - Sanity Check
Ralph,
Thanks for the suggestion. That sounds like an awful lot of work, though.
Steve
On Wed, Apr 4, 2012 at 12:55 PM, Ralph Mitchell <user-00a5e44c48c0@xymon.invalid>
wrote:
Steve,
On your side-note - I needed to do pretty much the same thing, for a
dog-n-pony presentation to management. I don't know the *best* way to
do it, but I got a second set of pages up by duplicating
/home/xymon/server and changing a bunch of references in
xymonserver.cfg in the copy to point to the copy structure. Then I
replicated the [xymongen] entry in the original
xymon/server/etc/tasks.cfg and pointed ENVFILE to the copy.
Some of the reports still pull up all the hosts, but the alternate
all-non-green page only shows systems that are listed in the
alternate's hosts.cfg. If you have your systems split out into
multiple files under hosts.d, you could just link the relevant file to
the copy to avoid duplication of effort.
I'm sure it can be done better, I just needed something *now* rather
than *perfect*...
As for ghost entries, I have a script that converts the ghost list
into an "Unconfigured Client" page so that any new system shows up
there within about 10 minutes of first checking in. People in other
groups were installing the client on a bunch of SuSE systems I don't
have access to, and we're also installing the client as part of a RHEL
kickstart from Satellite.
Ralph Mitchell
On Wed, Apr 4, 2012 at 11:59 AM, Steve Holmes <user-ec1bf77b1b44@xymon.invalid> wrote:
Don,
We have wrestled with the same issues. We started with systems organized
by
OS (Unix/Windows) and then as more apps became multi-platform have moved
away from the platform centric organization, with some exceptions. The
reason for the change is so we can see at a glance when there is a
problem
in a service we support so when there is a problem the customers for
that
service can be notified, unless the problem is fixed before the
customers
have to be notified (which is the big payoff with using Xymon).
Our main page contains 3 groups:
Services
Platform Support
Infrastructure
Under Services there are sub pages:
Production
Non-Production
Pre-production
Decommissioned
Under Platform Support there is currently only:
Platform Windows Servers
Under Infrastructure:
Authentication
Network
Server Provisioning
Prod and non-prod each have a list of application/service areas as sub
pages, each of which is a list of hosts in logical groups with no
respect
for OS platform. Within the groups the hosts are listed in alpha order.
Pre-production contains hosts which are not in production yet, but will
be
heading there (with some arm twisting at times). The reason for this is
the
OPS center only calls support for alerts that show up on a production
page.
Hosts in pre-prod (as well as non-prod) can fail without causing a call.
Decommissioned is where we put host entries for hosts that are just
that. We
keep them there for a year after they've gone off line in case someone
wants
to see the history. They all have noconn and all the NOPROPS so they
don't
show up anywhere else.
The Infrastructure group is also production, but not application
specific.
This is an area currently under development so it is incomplete. There
we
have network devices, DNS servers, and the like.
Platform Support was a special request from the Windows admins to group
all
of the windows servers in one place (with duplicate entries) so they
don't
have to look through all of the application pages to find their servers.
The
Platform Windows Servers sub page contains sub pages for Prod and
Non-Prod,
each of which is grouped by application area. Yes, this duplicates the
work
I have to do when Windows systems are added, but they know that if they
don't tell me exactly where to put the duplicate entry it won't go in.
We
could also put a page in there for Linux/Solaris admins, but that hasn't
been requested, yet.
Many times when a new server shows up in the ghost report I have to ask
the
admins for information about where it should go. Our naming convention
helps, but not totally.
Side note: OPS likes to watch the all-non-green page. But that contains
non-green tests for non-prod as well as prod. I would really like to be
able
to provide them with an all-non-green-prod-only (for lack of better
terminology) so they could easily see what they need to. Putting NOPROPS
on
all non-prod would prevent the admins from being able to use the same
page
to watch everything. Something I'm not willing to do.
HTH
Steve
On Wed, Apr 4, 2012 at 10:57 AM, Don Kuhlman <user-5eb2bfadc6c6@xymon.invalid>
wrote:
Hi folks. I have been modifying our xymon server host cfg file setups.
I
have been moving page layouts around. I thought I would send a note to
the
list to see what others are doing in their web page layouts just to
have a
sanity check…
Do you set up your main page to list things by OS, then by environment
–
like this:
Unix - then Prod, Dev, Test, Uat, etc.
Windows – then Prod, Dev, Test, Uat, etc.
Do you also use Application groups and then arrange them by OS and
environment ?
App1, Unix, Prod
App1, Unix, Dev
Or
App1, Prod
App1, Dev
Here's what I've been doing and I'm having second thoughts about the
logic
of doing it this way:
Main xymon page lists the following Pages
Server lists by hostname Applications Infrastructure Other Systems
Under Server lists by hostname – I have now made up UNIX-MAC and
WINDOWS
Under each of these I have PROD and DEV
Under the Applications I have several business Applications -
App1
App2
App3
In each of the App1, App2, App3, I have Prod and Dev subpages
I'm creating include files for each category – like HostsApp1Prod.cfg,
HostsApp1Dev.cfg, HostsApp2Prod.cfg, HostsApp2Dev.cfg, etc.
Now that I've changed it, I will probably need to create new
HostsApp1ProdUnixMac.cfg, HostsApp1ProdWindows.cfg
I would like to be able to setup base rules for monitoring the Prod &
Dev
systems – Prod disk, mem, cpu is different than Dev disk, mem, cpu,
etc.
That's why I thought breaking out by OS and then environment would
make
sense.
Then I want to create very specific service, process, or other
monitoring
for the application servers.
Does this seem like a good way to go, or am I making it too complicated
by
breaking everything down this way?
Thanks
Don K
--
If they give you ruled paper, write the other way. -Juan Ramon Jimenez,
poet, Nobel Prize in literature (1881-1958)
I prayed for freedom for twenty years, but received no answer until I
prayed
with my legs. -Frederick Douglass, Former slave, abolitionist, editor,
and
orator (1817-1895)
--
If they give you ruled paper, write the other way. -Juan Ramon Jimenez,
poet, Nobel Prize in literature (1881-1958)
I prayed for freedom for twenty years, but received no answer until I
prayed with my legs. -Frederick Douglass, Former slave, abolitionist,
editor, and orator (1817-1895)