Firestore is a NoSQL document database built for automatic scaling, high performance, and ease of application development. It is the newest version of Datastore and introduces several improvements over Datastore.
Because Firestore in Datastore mode is optimized for server use cases and for App Engine, we recommend using Firestore in Datastore mode for databases that will be used primarily by App Engine apps. Firestore in Native mode is most useful for mobile and real-time notification use cases. For more information about Firestore modes, see Choosing between Native Mode and Datastore mode.
Using Datastore mode with App Engine
To use Datastore mode with App Engine:
If you haven't already done so, create a database and choose Firestore in Datastore mode.
You can use existing Datastore databases with App Engine apps. These existing databases will be automatically upgraded to Firestore in Datastore mode.
New Python 3 apps should use the Datastore mode client library to interact with Datastore mode:
If you are planning to upgrade your Python 2 app to Python 3, you have a few options:
If your Python 2 app uses App Engine NDB to interact with Datastore, we recommend migrating to Cloud NDB instead of the Datastore mode client library as the former provides a much more similar user experience.
You can also choose to port your Python 2 app that uses App Engine NDB to Python 3 first and then migrate to Cloud NDB. To learn how to access App Engine NDB in Python 3, see the Accessing App Engine bundled services for Python 3 page.
For additional resources plus an example of migrating from a Python 2
app using App Engine NDB to an equivalent Python 3 Flask app with App Engine NDB
enabled, see the bundled services for Python 3
Datastore mode uses indexes for every query your application makes. The indexes are updated whenever an entity changes, so the results can be returned quickly when the app makes a query.
Datastore mode automatically creates single-property indexes for use with
simple types of queries. For complex queries that include multiple properties,
you'll need to configure composite indexes in your app's
The App Engine development server will update your
with the composite indexes needed to run your tests. Similarly, the
Datastore mode emulator can generate indexes
when you run tests.
You can also add the indexes to your app's
index.yaml file manually if you do not run local
tests or your tests do not include complex queries.
Setting database permissions
By default, your app has all the permissions required to read and write to Datastore mode and Firestore databases in your Google Cloud project.
To manage these permissions, each App Engine app uses a default service account that gives full read and write access to Datastore mode and Firestore databases in the same project as the app. You can change the permissions of the default service account, but your app may lose access unless you assign an IAM role with the required permissions.
For information about allowing other users, apps, or projects to access a database, see Accessing your database.
Using the Datastore mode emulator for local testingThe Google Cloud CLI includes a local emulator of the production Datastore mode mode environment. You can use the emulator to develop and test your application locally. In addition, the emulator can help you generate indexes for your production Datastore mode mode instance and delete unneeded indexes.
If you use the App Engine local development
server to test your
app, you can ensure that the server will use the Datastore mode
emulator by setting the
--support_datastore_emulator=true flag when starting
If you are using the Datastore mode emulator, dev_appserver will display:
... Using Cloud Datastore Emulator.
Pricing, quotas, and limits
Datastore mode offers a free quota with daily limits. Paid accounts offer unlimited storage, read, and write operations. More information is available on the Datastore Quotas page.