Databricks Lakebase connections

Databricks Lakebase is a PostgreSQL-compatible OLTP database. DataMasque connects to it as a standard database connection and masks the data in place.

• Required privileges
• Encrypted connections
• Limitations
• Supported data types
• Unsupported key columns

Required privileges

Databricks Lakebase is PostgreSQL-compatible. A DataMasque connection for Databricks Lakebase requires the same privileges as a PostgreSQL connection. Please refer to PostgreSQL connections for the required privilege statements.

Databricks Lakebase requires native PostgreSQL username and password authentication. These credentials are long-lived and do not expire.

Note: OAuth-based authentication is not supported for Databricks Lakebase connections in this version.

Databricks Lakebase is a managed service with the following platform constraints that are relevant when planning masking jobs:

• No superuser access. DataMasque does not require superuser privileges on Lakebase. All masking operations use standard database privileges.
• No tablespaces, logical replication, or event triggers. DataMasque does not use any of these features, so no special configuration is needed.
• Curated extension list. Lakebase provides a subset of PostgreSQL extensions. DataMasque uses only extensions available on the platform.
• Connection lifetime and idle timeout. Lakebase enforces a 3-day maximum connection lifetime and a 24-hour idle timeout. See Databricks Lakebase Limitations for the impact on masking runs.

Encrypted connections

Databricks Lakebase requires SSL for all connections. DataMasque always connects with sslmode=require. No additional configuration is required in DataMasque.

Limitations

• Expression and functional indexes are not exposed via pg_catalog.pg_index on Databricks Lakebase and are therefore not detected or managed by the auto_manage_indexes feature. Standard BTREE, GIN, and partial indexes are managed as normal. Masking still completes successfully when expression indexes are present; they are simply left in place during the masking run.

• Databricks Lakebase enforces a maximum connection lifetime of 3 days and an idle timeout of 24 hours. These limits are imposed by the platform and are unaffected by DataMasque configuration. Masking jobs that run longer than 3 days may fail due to connection termination; consider splitting very long-running jobs into smaller batches.

Supported data types

Databricks Lakebase is PostgreSQL-compatible and supports the same data types as PostgreSQL. See PostgreSQL data types for the list of supported types.

Unsupported key columns

Databricks Lakebase has the same unsupported key column types as PostgreSQL. See PostgreSQL connections -> Unsupported key columns.