Hello,
As much as I think it's a silly idea:
A company has a policy stating that all production systems need to be
monitored by a Microsoft SCOM system. Now, there are servers and hosts
that can only be monitored by Nagios (e.g. HPC clusters where the compute
nodes are hidden on isolated
On 20 December 2011 13:58, Troels Arvin tro...@arvin.dk wrote:
Thoughts/comments?
In the other direction, we have an IBM Director system which simply
forwards alerts to Nagios as SNMP Traps. It seems to work pretty well
on the whole.
I would think you could configure your Nagios system to
Troels,
I'm doing what Jim suggested by sending Nagios traps to an IBM Tivoli Netcool
system and it's working fine. I only had to issues that have bitten me enough
to make me look at an approach similar to what you are suggesting in your ideas.
The first problem is the Check_NRPE tests will
Thanks to Jim and Mike.
The idea of Nagios sending traps actually is much in line with the SCOM
way of doing things, as far as I have understood.
Personally, I find the trap-based approach rather fragile, though:
- The receiver (SCOM) may be down when a trap is sent. Actually
not that far
There is a commercial Nagios to SCOM connector, we use it at my place
of work - avoid it and work on a custom solution (which is the path
you are on) - the commercial connector is awful
* Only allows for one Nagios system to send to SCOM using SNMP traps
* Does not translate statuses properly
*