[Xymon] Dropping Tests and Hosts

[Vernon Everett]

Hi all

I have seen a few posts recently regarding the misery of dropped tests, and
wanting the data back.
It's easy to understand how this happens. Often, management and/or support
groups are sure they don't need something, and then change their mind.
Not only is it a pain, but many times, the data is lost.
Here's how we tackled the problem.

We created a page off the main menu, called "Decommissioned", but what to
call it is largely irrelevant.
Any time a server was decommissioned, it would be moved to there, with the
NOPROP* tags. (See the hosts.cfg man pages)
This would allow us to go back and review any historical data if we chose
to.
In the event where the host name was recycled, I would use the rename
facility to rename to hostname.old, or something similar.
I also put a comment in the hosts.cfg, to indicate the date it was put into
the decommissioned page.
After it had been there for a (indeterminate) period of time, we would
eventually remove it and give it a decent burial.

The same strategy would be applied to dropped tests.
In the Decommissioned section, you have a second entry for your host, with
the "dropped" tests.
In the main page, use group-except, or one of the other group-* settings to
exclude the dropped test from view.
This would keep the dropped test alive, and possibly even collecting data,
but it would appear to be removed.

The advantage of this is that decommissioned servers are not dropped, and
dropped tests are not deleted, giving instant access to the data should we
need it.
Historical data is available for comparison, and old tests can be reinstated
quickly should somebody decide they really do need the data after all.

Not much help for you poor folk who have already lost data, but might help
preventing a similar event in the future.

Regards
      Vernon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xymon.com/pipermail/xymon/attachments/20111003/a2f95696/attachment.html>