This is a multipart message in MIME format.
--=_alternative 004EC13885256D09_=
Content-Type: text/plain; charset="US-ASCII"
Ethan,
I haven't really been following the discussion on 2.0 that closely, but it
seems like you have the feature set pretty mapped out already. I have a
feature suggestion that I hope you can consider for the following major
release if you don't already have plans to implement something similar.
There are several things I am currently trying to do with Nagios that I
cannot presently do easily. First of all, as Netsaint grew to Nagios more
and more distributed processing tools and logic were added however one
major feature was left untouched and that is multi-test consolidation.
Let's say I'm monitoring a Web site from 3 geographic locations using any
of the current methods to report the results to central Nagios server. I'm
interested in the availability of the Web site and not the network links
of any of the monitoring locations, so I don't want to be notified if one
or two probes can't reach the web site, however, if all 3 are down, I want
to send out notifications.
What I envision as being a solution to this issue would be to have a
"meta-service" record (similar to the current dependancy record) where one
could specify several services to group together. I would love to have
each of the results of each service available as macro variables and be
able to perform RPN (reverse polish notation) math on them to determine
the "meta-service" state (crit, ok, etc...) and perhaps a perl expression
to return a result string for the "meta-service".
This would allow a lot of flexibility in being able to notify under very
specific conditions. Haven't we all wanted to retrieve an SNMP OID,
compare it to another OID (or do some math on it) and then determine if we
need to alarm? Similarly, post collection processing will allow much
better integration with graphing tools such as RRDTool.
I would be interested to know what everyone thinks about this.
Daniel Koffler
dan@julieanddan.com
Tel: 514.497.1411
Fax: 206.600.4642
GPG Key ID: 0xA2C6DC83 Fingerprint: 1FD7 3FDF 8A0D 961F 26A2 3EDA AE8F
A874 A2C6 DC83
--=_alternative 004EC13885256D09_=
Content-Type: text/html; charset="US-ASCII"
Ethan,
I haven't really been following the
discussion on 2.0 that closely, but it seems like you have the feature
set pretty mapped out already. I have a feature suggestion that I hope
you can consider for the following major release if you don't already have
plans to implement something similar.
There are several things I am currently
trying to do with Nagios that I cannot presently do easily. First of all,
as Netsaint grew to Nagios more and more distributed processing tools and
logic were added however one major feature was left untouched and that
is multi-test consolidation.
Let's say I'm monitoring a Web site
from 3 geographic locations using any of the current methods to report
the results to central Nagios server. I'm interested in the availability
of the Web site and not the network links of any of the monitoring locations,
so I don't want to be notified if one or two probes can't reach the web
site, however, if all 3 are down, I want to send out notifications.
What I envision as being a solution
to this issue would be to have a "meta-service" record (similar
to the current dependancy record) where one could specify several services
to group together. I would love to have each of the results of each service
available as macro variables and be able to perform RPN (reverse polish
notation) math on them to determine the "meta-service" state
(crit, ok, etc...) and perhaps a perl expression to return a result string
for the "meta-service".
This would allow a lot of flexibility
in being able to notify under very specific cond
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: an@julieanddan.co