Coming from Another NMS

Switching monitoring tools doesn’t mean starting from zero. If you already track your devices somewhere — SolarWinds, PRTG, LibreNMS, Zabbix, Nagios, or a spreadsheet you keep by hand — you don’t have to re-discover everything. Export that list, shape it into a simple CSV, and import it into GridNMS. Your devices land in inventory with the right names, classes, and sites, and GridNMS starts polling them on the next cycle.

You can still run Network Discovery afterward to fill in anything your old tool missed — but a CSV import is the fastest way to bring a known inventory across in one shot.

The CSV format

The importer reads a plain CSV with a header row. Each row is one device. Most columns are optional; the ones GridNMS uses to place a device are name, class, and site.

Column	Required	What it means
address	For monitored devices	The IP address or hostname GridNMS uses to reach the device. Leave it blank for an inventory record you don’t want monitored yet.
name	Yes	The device’s display name — usually its hostname.
class	Yes	The device class, e.g. Router, Switch, Firewall. Must match an existing class in your GridNMS.
site	Yes	The site (location) the device belongs to. Must already exist.
location	No	A free-text location note, e.g. a rack or closet.
collector	No	The name of the collector that should poll this device. Leave blank to let GridNMS pick one automatically.
tags	No	Labels for grouping and filtering, separated by semicolons, e.g. core;east;critical.
snmp_version	No	2c or 3. Sets which SNMP columns below apply.
snmp_community	No	The SNMP v2c community string.
snmp_user	No	The SNMP v3 username.
snmp_authkey	No	The SNMP v3 authentication key.
snmp_privkey	No	The SNMP v3 privacy key.

Example

address,name,class,site,location,collector,tags,snmp_version,snmp_community
10.20.0.1,core-rtr-01,Router,HQ,Rack A1,,core;critical,2c,public
10.20.0.2,dist-sw-01,Switch,HQ,Rack A2,,access,2c,public
10.20.1.5,edge-fw-01,Firewall,Branch East,,east-collector,edge;critical,3,
,spare-sw-07,Switch,HQ,Storeroom,,spare,,

The last row has no address, so it’s brought in as an unmonitored record — a placeholder you can fill in and turn on later.

Note

Column order doesn’t matter — GridNMS reads columns by their header names. Extra columns your old tool exported are simply ignored, so you don’t have to strip them out first.

How to import

1. Open the Devices page and choose Import CSV.
2. Pick your file. GridNMS reads it and shows a preview before anything is created.
3. Review the preview. Each row is sorted into one of three buckets:
   • Will be created — a valid new device.
   • Skipped — an address already in your inventory (duplicates aren’t added twice).
   • Error — something’s wrong with the row, shown with a reason so you know exactly what to fix.
4. When the preview looks right, confirm the import. The devices that were marked “will be created” are added, and monitoring begins shortly after.

Tip

Bad rows are always reported, never silently dropped. If some rows have errors, fix them in your CSV and re-import — GridNMS skips the devices it already created and only adds the ones it couldn’t before.

Exporting from your current tool

Almost every monitoring tool can produce a list of the devices it watches. The exact steps differ by product, but the general idea is the same everywhere:

• SolarWinds — export a device/node list to CSV (or run a report over your nodes).
• PRTG — export your device tree or a device list.
• LibreNMS — export the device list from the UI or query it via the API.
• Zabbix — export hosts, or pull the host list via the API.
• Nagios — your host definitions already list every host; extract name and address from them.

Whatever you export, the job is to map its columns onto the ones above: find the fields that correspond to name, IP/hostname, device type, and location, rename the headers to match GridNMS (name, address, class, site, …), and drop or ignore the rest. A quick pass in a spreadsheet is usually all it takes.

After the import

• Monitoring starts automatically. Imported devices begin polling on the next cycle. Devices with valid SNMP settings start reporting metrics; devices without credentials are still checked for reachability (Up/Down).
• Set credentials by class, not per device. Rather than repeating an SNMP community on every row, set monitoring credentials once and let them apply across a class. Override only the handful of devices that are different.
• Bulk-edit to finish organizing. On the Devices page, select many rows at once to assign a class, site, tags, or collector to all of them in one action — handy for tidying up after a large import.
• Export any time. GridNMS can export your inventory back to CSV, so you can round-trip your list, keep a backup, or move devices between environments.

Troubleshooting

What you see	What it means	How to fix it
Class not found	The class value doesn’t match any device class in GridNMS.	Create the class first, or correct the spelling in your CSV to match an existing one.
Site not found	The site value doesn’t match an existing site.	Create the site first, or fix the name in your CSV.
Address skipped	A device with that address is already in inventory.	Nothing to do — GridNMS won’t create a duplicate. Remove the row if you want a clean preview.
Unmonitored record	The row has a blank address.	Expected. It’s added as an inventory-only placeholder; fill in the address and turn monitoring on when you’re ready.

Where to go next

• Manage what you imported in Devices & Inventory.
• Catch anything the export missed with Network Discovery.
• Set up polling credentials and packs in Monitoring.