When your data flow uses stateful processors such as ListSFTP, you can store
processor state in Firestore so that state persists across function invocations. Setting up
the Firestore integration helps you avoid losing track of processed files or other state
between function executions.
By default, if your data flow contains any stateful processors
(e.g. ListSFTP), this state is automatically stored in a Firestore collection called
nifi_state. You can set this name using the FIRESTORE_STATE_COLLECTION
environment variable. In order to integrate with Firestore, some additional steps are
required. Without these steps, the data flow will still run, but the state will not be
preserved.
In the GCP console, navigate to the Firestore service.
Click Select Native Mode.
Select the appropriate region, and click Create database.
Navigate to the IAM service, and click the Edit (pencil) button
on the service account principal that runs your Cloud Function.
Click Add Role, and add the Firebase
Admin role.
Click Save.
The Firestore state provider can be disabled even if your data flow contains
stateful processors by setting the DISABLE_STATE_PROVIDER Environment
Variable to true.
