Skip to content

Adding Assets Manually

Most of your inventory arrives on its own — sensors find it, cloud connectors pull it, a spreadsheet or a CMDB export brings it in. But some things you simply know about: an appliance nothing can scan, a business service that has no address at all, a host that is behind a firewall your sensors cannot reach.

Inventory → All assets → New asset is where you record those.

Where the button is

Open Inventory in the left rail. On All assets — and on the Certificates, Keys, Configuration, TLS, SSH, Data Protection, 3rd Party and Stale lenses too — the New asset button sits at the top right of the page.

You need the create assets permission to see it. Without that permission the button is not shown at all, rather than shown and then refused.

Pick the class first

The form opens on a single field: Class.

That order is deliberate. An asset's class is what kind of thing it is — Server, Switch, Virtual machine, Object storage, Business service — and the class decides which fields the rest of the form offers you. A server has an operating system and a serial number; an object store has a provider, a region and a bucket name. Neither list would make sense on the other.

The picker is the whole class taxonomy as one indented list, so you can see where a class sits in the tree while you choose it. Pick the most specific class that is true. If you are not sure, a parent class is a better answer than a wrong child — you can reclassify later, and the asset's History tab records that you did.

Once you pick one, a short description of that class appears under the field, and the Class attributes section fills in with the fields that class declares. Leave any of them blank; blank means "not recorded", which is a different and more honest thing than an empty value.

The rest of the form

Section What goes in it
Display name What a person calls this thing. Optional — the platform derives one from the identifiers if you leave it empty. For a service class it is required, because a service has no address or serial to be known by: its name is its identity.
Identifiers How this thing is known. See below — this is the important one.
Class attributes Whatever the class declares. Some classes declare none, and the form says so instead of showing an empty box.
Context Environment, support group, business unit, owner email. "Who do I page about this?" is usually the first question anyone asks of an asset, so there is a field for the answer.
Description One line, free text.
Tags Key/value pairs of your own.
Metadata An optional JSON object for anything the class schema does not declare.

Which identifiers you may type

An identifier is a fact that says which real thing this record is about. They are what let two sightings of one machine become one asset instead of two — so the form will not let you save a new asset with none at all. (The one exception is a declared service, which identifies by its name.)

You may enter:

  • FQDN, Hostname, IP address, MAC address
  • Serial number
  • SSH host key fingerprint
  • CMDB sys_id
  • Name (for service classes)

Two kinds are minted by a collector and can never be typed:

  • Agent ID — issued by the platform to a running sensor or agent
  • Cloud resource ID — the provider's own name for a resource

Both mean "this is the thing that agent is" or "this is that exact cloud resource", and only the platform can know either. If a person could type one, they could claim an identity the platform assigns — which is how two genuinely different assets get merged by hand, with no proposal and nothing to review.

A serial number you type is worth exactly as much as one a scanner read. What differs is the source, which the platform records against the identifier either way.

Editing an existing asset

When you reopen the form on an asset that already exists, you will see identifier rows with a padlock beside them. Those were recorded by a collector, not typed by a person, and the form will not retire them. They are shown rather than hidden: an asset that is matched on an agent ID the form never mentioned is confusing in a way that showing the locked row fixes.

If you remove an editable identifier and save, and the platform keeps it anyway, it tells you which one and why rather than reporting a silent success.

What happens when you save

The asset is not simply inserted. It goes through the same identification engine every discovery goes through, which asks first whether this is something you already have:

  • No match — a new asset is created.
  • Match — your entry is folded into the existing asset, which gains whatever identifiers, attributes and context you supplied. This is the point of running a manual entry through the engine: typing in a host a sensor already found updates that host rather than creating a duplicate beside it.
  • Conflict — your identifiers belong to more than one existing asset and nothing can settle which. Nothing is created or changed; a merge proposal is opened and the form tells you to review it in Discovery → Approvals.

Does it go straight into inventory, or into Approvals?

That is decided by your own approval policy, not by the fact that you typed it.

An asset — however it arrives — joins active inventory immediately only when it falls inside a prefix you defined with auto-approve turned on (Inventory → Networks). Anything else waits in Discovery → Approvals for a person.

So in practice:

  • A host inside an auto-approving prefix: monitored straight away, and it shows up in All assets on your next refresh.
  • Everything else — including anything with no address to match a prefix against, such as a business service or an object store: pending approval. It will not appear in the asset list until you accept it, and All assets shows a banner with the pending count and a link into the queue.

This is on purpose. The approval policy is the tenant's, and letting a hand-entered asset step around it would make the policy advisory. If you regularly add assets by hand and want them monitored immediately, define the prefix they live in and mark it auto-approve.

An asset in conflict always waits for a human, whatever the prefix rule says — there is a merge to settle first.

Certificates are a separate button

A certificate is not an asset, and it has its own path in: Add certificate, beside New asset on the Certificates lens. It opens a dialog with two tabs.

  • Upload PEM — paste a PEM or pick a file (leaf first; a chain is fine). The certificate's contents are read out of the PEM rather than typed, so nothing about it is taken on trust.
  • Declare — for a certificate you know exists but cannot upload, give its common name and expiry date and the platform alerts you before it expires. See Certificate monitoring.

Either way the certificate appears in the Certificates lens straight away, shown as unassigned until something links it to an asset.

Adding a certificate needs the manage assets permission.