You are browsing documentation for a version other than the latest stable release. Switch to the latest stable release, 2.3.
Unlike Device identity attributes, Device inventory attributes do not need to be unique. They simply provide information about the target device.
The Mender client will periodically collect inventory-related data for reporting to the Mender server. Executable files, stored in a known location, will be invoked by the Mender client, producing the inventory data as a set of key-value pairs.
When you log in to mender server UI and navigate to the device list, among other items, you can see certain pieces of data associated with each device. Those can be a hostname, IP address, or any key-value pair you can think of. For further details, please refer to the client configuration section.
The Mender server stores information about each device in the form of key-value pairs called attributes grouped in scopes. The inventory attributes belong to inventory scope which also denotes the ownership of data; the device "owns" the inventory data in a sense that it can overwrite the attributes in the scope at will.
Some considerations while working with inventory attributes:
The following screenshot shows an example set of inventory data in the expanded view of a device.
The inventory attributes mechanism allows you to publish to the Mender server every piece of data that is important for your usage and that can be expressed as key-value pairs. Later you can use these attributes to create filters, dynamic groups, and to search for devices.
Examples use cases:
Adding more inventory fields, such as the kernel version or total available memory, is easy to do and Mender supports it well. Then you can use the inventory fields to create Dynamic groups (see the screenshot below).
For more information on customizing the Inventory attributes, please refer to the client configuration section.