Measurements With the SDK

Measurements With the SDK

Use the current released SDK supported by your Custd deployment. Initialize client with your tenant's authenticated base URL and credential flow before these snippets; see SDK contract. These are workflow examples after initialization, not standalone programs. Replace historical target/observation dates with the intended measurement period.

TypeScript

await client.admin.measurement.projects.create({
  projectSlug: "checkout-runway",
  name: "Checkout Runway",
  kind: "deadline_forecast",
  series: [{
    seriesSlug: "checkout-completions",
    name: "Checkout completions",
    unit: "count",
    completionDirection: "increase",
    source: "manual",
  }],
  target: {
    targetSlug: "release",
    name: "Release",
    targetValue: 100,
    targetDate: "2026-08-31T00:00:00Z",
    state: "active",
  },
});

const csv = "seriesSlug,observedAt,value\ncheckout-completions,2026-07-01T00:00:00Z,42\n";
await client.admin.measurement.projects.importCSVString("checkout-runway", csv, 1);

Go

project, err := client.Admin.Measurement.Projects.Create(ctx, custd.MeasurementProjectCreate{
	ProjectSlug: "checkout-runway",
	Name:        "Checkout Runway",
	Kind:        "deadline_forecast",
	Series: []custd.MeasurementSeriesCreate{{
		SeriesSlug:          "checkout-completions",
		Name:                "Checkout completions",
		Unit:                "count",
		CompletionDirection: "increase",
		Source:              "manual",
	}},
	Target: custd.MeasurementTargetCreate{
		TargetSlug:  "release",
		Name:        "Release",
		TargetValue: 100,
		TargetDate:  "2026-08-31T00:00:00Z",
		State:       "active",
	},
})

PHP

$project = $client->adminMeasurementProjects()->create([
    'projectSlug' => 'checkout-runway',
    'name' => 'Checkout Runway',
    'kind' => 'deadline_forecast',
    'series' => [[
        'seriesSlug' => 'checkout-completions',
        'name' => 'Checkout completions',
        'unit' => 'count',
        'completionDirection' => 'increase',
        'source' => 'manual',
    ]],
    'target' => [
        'targetSlug' => 'release',
        'name' => 'Release',
        'targetValue' => 100,
        'targetDate' => '2026-08-31T00:00:00Z',
        'state' => 'active',
    ],
]);

Python

project = client.admin.measurement.projects.create({
    "projectSlug": "checkout-runway",
    "name": "Checkout Runway",
    "kind": "deadline_forecast",
    "series": [{
        "seriesSlug": "checkout-completions",
        "name": "Checkout completions",
        "unit": "count",
        "completionDirection": "increase",
        "source": "manual",
    }],
    "target": {
        "targetSlug": "release",
        "name": "Release",
        "targetValue": 100,
        "targetDate": "2026-08-31T00:00:00Z",
        "state": "active",
    },
})

SDK helpers validate that import responses contain one result per submitted row and that every successful row includes observationUuid. Failed rows retain their RFC 9457 problem fields.

Keep the created project's identifiers, import one valid row and inspect its result before adding volume. Then run a forecast and inspect Measure. A project without observations is not a populated forecast. Handle duplicate slugs by inspecting the existing project; do not blindly repeat creation after an uncertain response.