[rancid] Monitoring rancid poll failures

[Austin Schutz]

I would like to use an alternative to email monitoring of rancid poll
failures. It would be nice if there were some way of displaying these
failures alongside other types of alarms.

I have been considering injecting alarms into netcool, which we have
locally, but I would like to hear what other people are using.

I have taken a brief glance also at nagios, zenoss, and opennms. But
none of these appear to offer much flexibility when you get past initial
stages. I don't want something to go out and do network polling, I just
want to be able to create my own alarms which the interface should do
intelligent things with.

Also I don't want them doing snmp/other polling of the network, because
I already have a system in place for that. I want to give <whatever>
system alarms, and have it do something reasonable with them.

Also, when these systems (nagios, zenoss, etc.) offer the ability to
generate an alarm in some custom fashion, it is typically possible to
set a single alarm when something goes wrong, but not to generate
multiple alarms at once. I would like to have one alarm for each device
which rancid can't poll.

In fact, I would like to not only send <whatever> rancid alarms, I would
also like to use it for catching alarms from other applications I
babysit/develop.

How are other people managing this?


Austin

[alex]

Post by Austin Schutz
I have taken a brief glance also at nagios, zenoss, and opennms. But
none of these appear to offer much flexibility when you get past initial
stages. I don't want something to go out and do network polling, I just
want to be able to create my own alarms which the interface should do
intelligent things with.

You might like to have a look at Zabbix. You can create a host with any number
of checks, which could be SNMP, simple TCP, or external script [amongst
others]. The external script zabbix uses to check a host could easily be
RANCID.

alexd

[Lance Vermilion]

Austin,

I am sure you already know RANCID is going to be an option with OpenNMS's
next stable release. Pretty handy but I think you were looking at a way to
add an API like concept to RANCID so other NMS tools could tie into it.

Post by alex

Post by Austin Schutz
I have taken a brief glance also at nagios, zenoss, and opennms. But
none of these appear to offer much flexibility when you get past initial
stages. I don't want something to go out and do network polling, I just
want to be able to create my own alarms which the interface should do
intelligent things with.

You might like to have a look at Zabbix. You can create a host with any number
of checks, which could be SNMP, simple TCP, or external script [amongst
others]. The external script zabbix uses to check a host could easily be
RANCID.
alexd
_______________________________________________
Rancid-discuss mailing list
http://www.shrubbery.net/mailman/listinfo.cgi/rancid-discuss

[Peter Klausner]

In a limited way, you can do that now - from your email configuration.
Simply set up a pipe as [one of] your mail recipient[s]. A script
then can do whatever magic. This sample entry in /etc/mail/aliases
creates syslog messages - notice for the normal output, err when
something goes wrong:

rancid-group: "|/bin/logger -p user.notice -t rancid.group"

rancid-admin-group: "|/bin/logger -p user.err -t rancid.group"


Peter

[Austin Schutz]

Post by Lance Vermilion
Austin,
I am sure you already know RANCID is going to be an option with
OpenNMS's next stable release. Pretty handy but I think you were looking
at a way to add an API like concept to RANCID so other NMS tools could
tie into it.

I wanted to check this out before actually responding. It is pretty cool
that there is a way to access the inventory through the nms itself. Not
what I had in mind at all, and fairly limited, but pretty cool.

I want the opposite of what you suggest. I want an API like concept in
my NMS tool so I can tie into that. I want something intuitive I can
generate my own alarms with, not provide an API from my tool (rancid in
this case) or hork together my own snmpd or trap generator to try to
make something appear in the nms.

I think I may be able to bend the send-event.pl script included with
opennms to my will. In fact, it already seems to understand that rancid
can have event failures, so that's humongous progress right there.

I'm also going to check out Zabbix.

Thanks for the tips. These are good tools, and certainly the price can't
be beat.

Austin