Data Models

View, version, and compare data models that describe a connector's available data points.

A data model is a structured representation of all data points available through a connector. It describes what data a device or server exposes — the names, hierarchical structure, data types, and addresses of every readable data point. Data models are used for match suggestions, address reference during manual binding, and understanding the scope of a data source.

The Data Model tab on the connector detail page is available for all connector protocols — not just OPC UA.

Viewing the data model

Open a connector detail page and go to the Data Model tab. The primary view shows the current data model — the latest version — as an interactive tree browser.

Tree browser

The tree view displays data points in a hierarchical structure, reflecting how the data source organizes its variables. You can expand and collapse nodes to navigate the hierarchy.

Each data point in the tree shows:

Name — the human-readable name of the data point
Path — the full hierarchical path (for example, WindFarm > Turbine_001 > ActivePower)
Data type — the value type (Float64, Int32, Boolean, String, and others)
Address — the protocol-specific identifier used in bindings (for example, ns=2;s=ActivePower for OPC UA)

The tree browser loads data lazily — only the children of expanded nodes are fetched. This keeps the view responsive even for data models with thousands of data points.

Data model summary

Above the tree browser, the data model summary shows:

Version number — the current data model version
Total data points — how many data points are in the model
Created date — when this version was created
Source — how the version was created (discovery crawl or manual upload)

Data model versions

Every time the data model is updated, a new version is created. Previous versions are preserved so you can track how a data source's structure has changed over time.

Version history

Click the version dropdown to see the full version history. Each entry shows:

Version number
Creation date
Total data points
Source type (discovery or upload)

Select a previous version to view its data model in the tree browser.

Comparing versions

To compare two versions, select a version and click Compare. The diff view shows:

Added data points — data points present in the newer version but not in the older one, highlighted in green
Removed data points — data points present in the older version but not in the newer one, highlighted in red
Changed data points — data points that exist in both versions but with different attributes (for example, a changed data type), highlighted in yellow

This is useful for understanding what changed when a device firmware was updated, a PLC program was modified, or an OPC UA server configuration changed.

Creating data model versions

Data model versions can be created in two ways: discovery crawl (for protocols that support server browsing, such as OPC UA) and manual upload (for all protocols).

Discovery crawl (OPC UA)

For OPC UA connectors, you can run a discovery crawl that connects to the server, traverses its address space, and saves all discovered data points as a new data model version.

To run a discovery crawl, expand the Run Discovery section on the Data Model tab and configure the crawl parameters:

Parameter	Description
Start Node	The root node to begin crawling from (for example, "Objects")
Max Depth	Maximum depth to traverse. Leave empty for unlimited.
Include Values	Whether to read current values from variable nodes during the crawl

Click Start Crawl to submit the job. The crawl runs asynchronously. You can track its progress in the Discovery History section below.

Discovery job statuses

Status	Description
Pending	Job is queued and waiting to start
Running	Actively crawling the server. Progress and node count are displayed.
Completed	Crawl finished successfully. A new data model version has been created.
Failed	An error occurred. Check the error message for details.
Cancelled	The job was cancelled by a user

When a crawl completes successfully, the platform automatically creates a new data model version from the discovered data points. The Data Model tab updates to show the new version.

Manual upload

For connectors that do not support live discovery (for example, Modbus TCP), or when you want to define the data model without connecting to the device, you can upload a data model manually.

The upload accepts a structured list of data points, each with a name, address, data type, and optional path for hierarchy. This is useful when you have a register map or variable list from the device documentation and want to import it into the platform.

Manual upload is available for all connector protocols, including OPC UA. This allows you to define data models for devices that are not yet online or accessible.

Auto-rematch on new versions

When a new data model version is created, existing match suggestions from the previous version are automatically expired. If auto-rematch is enabled on the connector, the platform also submits a new matching job automatically, generating fresh suggestions based on the updated data model.

This ensures that your binding suggestions always reflect the latest state of the data source.

Using data models with bindings

The data model serves as the reference for creating bindings:

Match suggestions — the AI matching engine compares your variables against the data model's data points to suggest bindings automatically
Manual binding — you can reference the data model to find the correct address for a specific data point before entering it in the binding editor

Standalone OPC UA Discovery

Fleetera also provides a standalone OPC UA Discovery tool accessible from Tools in the sidebar. This tool lets you browse any OPC UA server by entering its endpoint URL, without needing to create a connector first. It is useful for initial server exploration before setting up connectors.

Data Models

On this page