Data model

The sandbox exposes two ways to publish data: metrics for named scalar values, and tables for lists of similar rows. Pick the one that matches the shape of the data you collect.

Metrics

Use metrics for a small, fixed set of named scalar values per device — “CPU temperature”, “uptime”, “firmware version”. Each metric has a stable identifier (UID), a label, a value, a unit, a type, and optional metadata.

var cpu = D.createMetric({
    uid: 'cpu',
    name: 'CPU Temperature',
    value: 62,
    unit: 'C',
    valueType: D.valueType.NUMBER,
});
var fw = D.createMetric({
    uid: 'fw',
    name: 'Firmware',
    value: '4.1.2-rc3',
    unit: '',
    valueType: D.valueType.STRING,
    metadata: { url: 'https://vendor.example/release-notes/4.1.2-rc3' },
});
D.success([cpu, fw]);

• The uid is the stable identifier the portal uses to align history across runs. Keep it short and stable.
• D.valueType.NUMBER values produce charts; D.valueType.STRING values produce a text field.
• metadata.url (optional) lets the portal render the metric as a link to an external resource.

Deprecated: D.createVariable

The positional D.createVariable(uid, name, value, unit, valueType) form is still available for backward compatibility but is deprecated. New drivers should use D.createMetric. See the API reference for migration details.

Tables

Use tables when the device returns a list of similar rows — interfaces, VLANs, processes, neighbours. Declare the columns up front, then add rows.

var ifaces = D.createTable('Interfaces', [
    { label: 'Interface' },
    { label: 'In',  unit: 'bps' },
    { label: 'Out', unit: 'bps' },
    { label: 'Up',  valueType: D.valueType.STRING },
]);

ifaces.insertRecord('eth0', ['eth0', 1200000, 850000, 'yes']);
ifaces.insertRecord('eth1', ['eth1', 0, 0, 'no']);

D.success(ifaces);

• The first argument to insertRecord is the row UID. Like variable UIDs, it should be stable across runs.
• Table cells can mix numbers and strings; use the column valueType to steer the portal renderer.

Choosing

You have	Use
A fixed handful of named values	Metrics
A single scalar value	Metric
A list of similar objects with the same fields	Table

You can mix metrics and a table in one D.success call when a device needs both — for example, device-level metrics plus a per-interface table:

D.success([fw, cpu], ifaces);

Next steps

• Quickstart — see metrics in action.
• API reference for D.createMetric.
• API reference for D.createTable.