Cloud Firestore Automatically Schedule Data Backup and Restore (Export and Import)

In the previous article, we learnt everything about how to export data from Cloud Firestore into Cloud Storage and import data from Cloud Storage into Cloud Firestore (vice-versa). A very common requirement is to have backup and restore operations scheduled to run at regular intervals. For instance backup the entire database daily without any manual intervention.

To schedule regular imports and exports, we could use the gcloud CLI utility or the Cloud SDK scripts (as shown in the previous article) with cron jobs. But there’s an even better way to do the same with a Firebase Function that is triggered by a Cloud Schedular job at regular intervals.

Export a function from your functions/index.js that looks like this:

const functions = require('firebase-functions');
const firestore = require('firebase-admin/firestore');

const client = new firestore.v1.FirestoreAdminClient();

exports.scheduledFirestoreExport = functions.pubsub
  .schedule('every 24 hours')
  .onRun(async (context) => {
    const projectId = process.env.GCP_PROJECT || process.env.GCLOUD_PROJECT;
    const databaseName = client.databasePath(projectId, '(default)');
    // Replace BUCKET_NAME with the actual value
    const bucket = `gs://[BUCKET_NAME]`;

    try {
      const responses = await client.exportDocuments({
        name: databaseName,
        outputUriPrefix: bucket,
        // Entire DB
        collectionIds: [],
      const res = responses[0];
      console.log(`Export Operation Response: ${}`);
    } catch (err) {
      throw new Error('Export Operation Failed');

Most of the export logic above is borrowed from the previous article. As far as the scheduling logic goes, we create a scheduled function using functions.pubsub.schedule('your schedule').onRun((context) => { ... }). In this case we schedule the function to run once every 24 hours (daily). This code creates a Pub/Sub topic and uses Cloud Scheduler to trigger events on that topic at the specified schedule that ends up running the function.

You may want to make the following changes in the code snippet above:

  • Change the BUCKET_NAME to the desired bucket name value.
  • Modify every 24 hours to something more suitable for your needs. You can use the unix-cron syntax or the AppEngine cron.yaml format.
  • If you want to export specific collection (groups), then specify them in collectionIds instead of an empty array.
  • If you want to import documents instead of exporting, feel free to borrow the Client SDK importDocuments code from the previous article.

Finally, deploy the scheduled Cloud Function code to Firebase:

# Deploy all functions
$ firebase deploy --only functions

# or Deploy just a single function
$ firebase deploy --only functions:scheduledFirestoreExport

Configure Appropriate Permissions

Your Cloud Function will use the project’s default service account to perform the import and export operations. The service account looks like this:

[email protected]

Performing import and export operations require permissions that are covered by the following two roles:

  • Storage Admin – The default service account should have this role’s permissions for the same project by default. So you need not do anything here.
  • Cloud Datastore Import Export Admin – You will have to make sure this role is assigned to the service account so that the required/relevant permissions are granted.

Trigger a Test Run

Once the function has been deployed, the scheduler job will be created which will trigger the Cloud Function at the specified interval. For testing though, you may want to trigger a test run. To do that, follow these steps:

Then to view the function logs:

Leave a Reply

Your email address will not be published. Required fields are marked *