Discovery or manual upkeep — which should I choose?

Requires Pro

Quick answer

All three approaches lead to assets in your inventory — they differ in depth of detail and effort. Manual upkeep (all plans) suits small, stable inventories and anything that isn’t a network device. The agentless scan (Pro) quickly gives you an overview of a network but only delivers basic data. The inventory agent (Pro) delivers the deepest data — including installed software — and keeps assets up to date automatically. In practice, you combine all three.

Comparing the methods

	Manual upkeep	Agentless scan	Inventory agent
Plan	all (Basic)	Pro	Pro
Setup	none	none (just specify a target)	token + script per device
Data captured	anything you enter	IP, hostname, open ports, OS guess, optional SNMP name	OS + version, manufacturer, model, serial number, installed software
Freshness	only when maintained	a snapshot per scan	automatic with every report
Duplicate protection	manual care needed	import status per device (409 on double import)	upsert via stable machine_id
Import into inventory	direct on creation	manual: import or ignore	automatic on first report
Coverage	anything (even furniture, contracts, peripherals)	only reachable network devices (max. 4096 hosts per scan)	only devices with Python 3.8+ and the agent
Role required	Technician+	Technician+	Administrator (token management)

Recommendations

• Start with a scan when you’re inventorying an existing network for the first time: it shows you within minutes what’s actually reachable. Import the relevant finds and ignore the rest.
• Roll out the agent wherever software inventories and details matter to you — typically on servers and managed clients. The agent keeps these assets up to date without further effort, but it does not automatically update the same assets you previously created manually or via a scan — it only recognizes its own devices (reported via machine_id).
• Maintain manually anything that isn’t an active network device: peripherals, licenses, certificates, decommissioned devices. Even after a scan or agent rollout, fields like location, department, or purchase dates remain manual upkeep.
• Repeat scans after network changes — every scan is a snapshot. New finds reappear with the status New; already-imported devices remain untouched.

What happens behind the scenes?

• Shared destination: All three approaches produce the same kind of assets with tenant assignment and an audit trail — reports, QR codes, and assignments work the same regardless of how the asset was captured.
• The asset limit applies everywhere: Whether manual, imported from a scan, or from a first agent report — every new asset counts against your plan’s limit (403 once reached).
• Separate sources: Scan finds and agent reports are not reconciled against each other. If you import a device from a scan and later install the agent on it, the first report creates a second asset (the agent finds no asset with its machine_id). In that case, remove the older scan-created asset, or skip the scan import for devices that will get the agent.

Related topics

• How do I start a network scan?
• How do I import discovered devices into my inventory?
• How do I install the inventory agent?
• How do I add an asset?
• Discovery — Overview