I'm not exactly sure of how to implement a solution to this problem but do have a comment on what the customer is doing.
I suspect that they have a procedure in place that whenever a new network connected asset comes online, it is ALWAYS configured to send traps to a standard trap destination. This is their procedural link between asset management and monitoring that they have come up with.
Pretty slick if you can get the network and server staff to follow the procedure.
From a monitoring point of view, this reduces the need to perform discovery and filter out the unwanted **** (desktops, laptops, printers, etc) and concentrate on things that matter for monitoring.
You could almost make the blanket statement that if the device doesn't sent a trap, then it doesn't need to be monitored.
yes, indeed. The plan is that they will have only one equipment type under the management system and don't want anything else on there.
If I ask the first part of the question again:
What happens when a trap arrives from an unknown device? Is it possible to trigger a script from this trap?
OK, it looks like I'm drawing a blank. I will assume that it's not possible to trigger a script when a trap arrives from an unknown device.
So, let's say I write a piece of code that will accept traps either on a different IP address or possibly on a different UDP port from Orion. This code will see the incoming trap and it must then create the device in Orion. It means that the factory configuration on the devices must be set to target these initial traps at this new application, not Orion. Is this creation of the device possible and how is it done? Do I use SQL or is there and API that will allow me to create the device?
Any other ideas on a way forward would be welcome. Failure is not an option here.