Backfill historical windows

How-to guideApplies to: Unified engine deployments

Preview historical schedule windows before creating runs. Use a saved cron or interval batch workload with data windows enabled and no incremental cursor.

Preview and start historical windows

Use an active daily midnight UTC workload with window filtering enabled. Inspect the preview before creating execution work.

Placeholder	Value
`{{workload_id}}`	Your environment value.
`{{revision}}`	`1`
`{{backfill_submission}}`	A new UUID for this request; reuse it when recovering the same submission.

Replace the placeholders in your client. Keep returned IDs for later requests. Paths below use the native API prefix; adjust it for your deployment.

1. Preview complete windows

POST /api/v1/workloads/{{workload_id}}/backfills/preview
Host: <ENGINE_HOST>
Content-Type: application/json

{
  "submission_id": "{{backfill_submission}}",
  "expected_revision": {{revision}},
  "from_unix_millis": 1767225600000,
  "to_unix_millis": 1767398400000,
  "parameters": {},
  "reprocess": "missing",
  "max_active_runs": 1
}

Endpoint reference.

2. Create the reviewed backfill

POST /api/v1/workloads/{{workload_id}}/backfills
Host: <ENGINE_HOST>
Content-Type: application/json

{
  "submission_id": "{{backfill_submission}}",
  "expected_revision": {{revision}},
  "from_unix_millis": 1767225600000,
  "to_unix_millis": 1767398400000,
  "parameters": {},
  "reprocess": "missing",
  "max_active_runs": 1
}

From the response, keep backfill_id from /id. Endpoint reference.

3. Inspect windows and run IDs

GET /api/v1/workloads/{{workload_id}}/backfills/{{backfill_id}}
Host: <ENGINE_HOST>

Repeat this read until /status reaches a terminal state: succeeded, failed, cancelled. Endpoint reference. Enter the workload ID and current revision under Example values. The example range covers January 1 and January 2, 2026, in UTC. It must align with the saved schedule’s complete windows. Use a daily midnight UTC schedule for these exact boundaries. Read every preview window and its exclusion reason. Missing, failed, and all have different reprocessing rules. Active or uncertain executions are not safe evidence for replay. Creating the backfill uses the reviewed request and a stable submission ID. Follow its window list and resulting run IDs. Cancelling a backfill stops further materialization and requests cancellation of submitted runs. Confirm final run states and commits. See Workload backfills for a complete inline-data workload and expected daily totals.

Transformation Engine

Learn Laminar

Workloads

SQL reference

HTTP API

Recipes

Operations

Backfill historical windows

Preview and start historical windows

1. Preview complete windows

2. Create the reviewed backfill

3. Inspect windows and run IDs

​Preview and start historical windows

​1. Preview complete windows

​2. Create the reviewed backfill

​3. Inspect windows and run IDs

Preview and start historical windows

1. Preview complete windows

2. Create the reviewed backfill

3. Inspect windows and run IDs