Retry failed dataflow runs

Support for retrying failed dataflow runs is available to batch sources. You can only retry dataflow runs that have failed.

This tutorial covers steps on how to retry failed dataflow runs using the Flow Service API.

Getting started

This tutorial requires you to have a working understanding of the following components of Adobe Experience Platform:

  • Sources: Experience Platform allows data to be ingested from various sources while providing you with the ability to structure, label, and enhance incoming data using Platform services.
  • Sandboxes: Experience Platform provides virtual sandboxes which partition a single Platform instance into separate virtual environments to help develop and evolve digital experience applications.

Using Platform APIs

For information on how to successfully make calls to Platform APIs, see the guide on getting started with Platform APIs.

Retry a failed dataflow run

To retry a failed dataflow run, make a POST request to the /runs endpoint while providing the run ID of your dataflow and the re-trigger operation as part of your query parameters.

API format

POST /runs/{RUN_ID}/action?op=re-trigger
The run ID that corresponds with the dataflow run that you want to retry.
An operation that determines the action to be performed. To retry a failed dataflow run, you must specify re-trigger as your operation.


You can use the re-trigger operation to retry successful dataflow runs as well, given that the successful dataflow run has zero ingested records.

The following request retries the dataflow run for run ID 4fb0418e-1804-45d6-8d56-dd51f05c0baf.

curl -X POST \
  '' \
  -H 'Authorization: Bearer {ACCESS_TOKEN}' \
  -H 'x-api-key: {API_KEY}' \
  -H 'x-gw-ims-org-id: {ORG_ID}' \
  -H 'x-sandbox-name: {SANDBOX_NAME}' \
  -H 'Content-Type: application/json'


A successful response returns a newly created flow run ID and its corresponding etag version.

    "id": "3fb0418e-1804-45d6-8d56-dd51f05c0baf",
    "etag": "\"1100c53e-0000-0200-0000-627138980000\""