We have a SAM Agent debate and we're about to install a lab to answer the debate I wanted to post this question on Thwack to get the expert advice.
Here's the debate:
How do agents report back to the polling engines, via polling engine DNS names or via polling engine IP address?
SAM Agent Settings:
Allow automatic agent registration - enabled
Automatically create node - enabled
Allow automatic agent updates - disabled
How long should newly registered agents be displayed as new - 24 hrs
Communication mode - Agent Initiated (Recommended)
Scenario:
We have a PCI-CDE environment that has firewall rules that don't allow inbound communication from the Solarwinds. Only outbound communication is allowed, and only one Polling Engine. This Polling Engine is designated as our PCI-CDE Polling Engine and agents are installed locally on the server by the server admins. PCI-CDE nodes cannot be moved to another Polling Engine.
We need to retire that existing PCI-CDE Polling Engine and replace it with a new one. The new server will have a the same IP address, but a new system name. The deal is we know and acknowledge a monitoring outage for the PCI-CDE environment. We want to mitigate the outage time window as much as possible, ease our workload and ensure all nodes are still being monitored in the end.
If the agents talk to the polling engines via IP address then we'll leave the agent settings as is, decommission the old sever and commission the new server with the same IP and in theory the agents will "perceptively" restore communication to the polling engine and won't care that it's a new server with a new system name.
Alternatively, if the agents care about the polling engine host name that we have to commission the new server with the old servers name and re-name it at a later point in time, which means extra steps.
So..... if anyone knows the answer it would help settle the debate and prevent diverted time in a lab installing a test environment.