Network Discovery Profiles - talk to me.

stuartd

We do use network sonar discoveries, but only ever for an initial scan in of devices.

What I'm interested in is in how people use these on an ongoing basis. 

I'm thinking about how you might manage new nodes being added - do you re-run the profiles on a regular basis and have a report advising what was added.. or ??

How do you manage any new interfaces or manage the profile so that it doesn't re-add interfaces you don't want to monitor, etc?

Any pros or cons?

TIA.

bobmarley

We do an ICMP only discovery of all of the subnets for each building, data center, etc once a month but do not import any of the nodes during discovery. We also have a polling engine set aside as a 'parking lot' for any new nodes that get added, so we can ensure the nodes get properly monitored before pushing them out to the production polling engines. 

Once all of the monthly discoveries are completed we use a powershell script which brings in the data from each discovery profile and does a lot of name matching on the nodes to determine which ones to import, then the script imports them as ICMP nodes to our 'parking lot' polling engine and assigns all of the custom properties for each node based on it's node type and the team that supports it.

Once the nodes get imported to the 'parking lot' we run a few other scripts to perform additional discoveries for WMI, SNMP, etc, based on the node type and then the nodes get moved over to the production polling engines. 

All of our alerts exclude alerting from any nodes in the 'parking lot' poller. 

For Servers, Timeclocks, and other devices that get added on a daily basis we don't use the discovery engine at all. We just use powershell scripts to pull the inventory from the external sources such as ServiceNow, VMWare, Timeclock Database etc. and the scripts do a compare against the SolarWinds database and any nodes that are present or missing get added or removed from SolarWinds. These nodes also get added originally as ICMP, the custom properties get set and then additional scripts run to do additional discoveries from powershell for each node type.

It was a lot of work to set this up but it ensures all of the nodes that need to be monitored get monitored.