Data Synchronization

Permify can connect to your databases to capture changes in your application’s database and sync authorization-related data in real time.

By using Debezium, Kafka, and the Permify Sync Service, your system can efficiently handle updates to permissions and roles without manual intervention.

​

Permify Sync Architecture Overview

​

1. Your Databases

The databases of your applications serve as the primary source, holding various data like user information, roles, permissions, and entity relationships. These databases contain the authorization-related data that will be captured and synchronized.

​

2. Debezium

Debezium is responsible for Change Data Capture (CDC), monitoring your application’s database for any changes (e.g., in tables like repository and organization). Debezium detects these changes in real-time and publishes them to Kafka, ensuring that any modifications are captured without impacting database performance.

​

3. Kafka

Kafka acts as the message broker, receiving change events from Debezium. It queues these events and passes them on for further processing.

​

4. Permify Sync Service

The Permify Sync Service is responsible for filtering the incoming data from Kafka. It identifies authorization-related changes (such as role assignments or permission updates) and synchronizes this data with Permify. Only relevant data is extracted and passed on.

• Role: Filters and processes the events from Kafka to focus on authorization data.
• Integration: Ensures that only the necessary authorization changes are synced.

​

5. Permify

Permify is an open-source authorization service. It applies authorization policies and ensures real-time enforcement of roles, permissions, and relationships based on the data synced from your databases.

• Role: Provides a centralized service to manage and enforce authorization policies.
• Functionality: Maintains up-to-date access controls, permissions, and entity relationships within your system, based on predefined policies.

​

Permify Sync Configuration

You can use the configuration file, which maps changes in database tables to Permify entities and their relationships.

By syncing the database changes with Permify, you ensure that any updates to the authorization data are handled in real time.

To give an example of how the configuration works, let’s examine the following schema.

​

Example Permify Schema

The Permify Schema defines the entities, relationships, and permissions that are relevant to your application. In this example:

• Entities like user, organization, and repository are defined.
• Relationships between entities, such as admin or member roles in an organization, are established.
• Permissions are specified based on relationships or attributes. For example, a repository can be viewed if it is public (is_public), and only admins or owners can edit it.

entity user {}

entity organization {
    relation admin @user
    relation member @user
}

entity repository {
    relation parent @organization
    relation owner  @user

    attribute is_public boolean

    permission view   = is_public
    permission edit   = parent.admin or owner
    permission delete = owner
}

​

Postgres Connector Example Config

In this example, the Postgres Connector is configured to track changes in the repository and organization tables:

• Entities: Each table is mapped to a Permify entity (e.g., repository, organization).
• Relationships: Columns like owner_id or organization_id in the database are mapped to relationships like owner or parent in Permify.
• Attributes: Columns such as is_public are mapped to entity attributes.
• Logging and Retry Policy: This configuration includes logging at the error level and a retry policy to handle connection issues with an exponential backoff strategy.
• Recovery: Backups are scheduled every 12 hours and stored in an S3 bucket.

# my_postgres_instance.yaml

connector: postgres
metadata:
  name: my_postgres_instance
tables:
  - name: repository
    permify_entity: repository
    primary_key: id
    relationships:
      - column: owner_id
        column_type: uint64
        permify_reference: owner
      - column: organization_id
        column_type: uint64
        permify_reference: parent
    attributes:
      - column: is_public
        column_type: boolean
        permify_reference: is_public
  - name: organization
    permify_entity: organization
    primary_key: id
    relationships:
      - column: user_id
        column_type: uint64
        join_table: organization_admins
        join_column: organization_id
        join_column_type: uint64
        permify_reference: admin
      - column: user_id
        column_type: uint64
        join_table: organization_members
        join_column: organization_id
        join_column_type: uint64
        permify_reference: member
logger:
  enabled: true
  level: error
retry_policy:
  max_retries: 5
  backoff_strategy: exponential
  backoff_interval: 2s
recovery:
  backup:
    enabled: true
    schedule: "0 */12 * * *"
    location: "s3://backup-bucket/my-postgres-instance-backups"

This setup ensures a seamless sync between your PostgreSQL database and Permify for managing authorization logic efficiently.

The Permify Sync Service is applied using a resource configuration, like the one above, through a CLI tool. You can apply this service using the following command:

permfiy-sync apply -f ./my_postgres_instance.yaml

This command ensures that the sync service is properly configured with your database.