lakebase-setup

Resolve Lakebase connection failures by verifying that your Databricks bundle includes the correct database or postgres resource in databricks.yml and that all required LAKEBASE environment variables are properly configured for your deployment type.

Postgres permission denied errors occur when the app service principal lacks explicit schema and table grants, which you can resolve by running the provided post-deploy permission grant script to prevent migration and checkpoint access failures.

Configure autoscaling Lakebase instances by adding the postgres resource to your databricks.yml file and wiring the required project, branch, and database ID environment variables to enable dynamic agent memory storage provisioning.

Missing embedding_dims errors arise from incomplete Lakebase environment variable configuration, requiring you to properly map the embedding dimensions in your Databricks bundle deployment to align with your long-term and short-term memory templates.

You can use either a provisioned Lakebase instance with a database resource or an autoscaling setup using a postgres resource in databricks.yml, depending on whether your agent chat history requires fixed capacity or dynamic project and branch scaling.

Prevent table schema migration errors by executing a critical post-deploy workflow that grants the Databricks app service principal explicit Postgres permissions for required schemas and tables before the agent begins writing long-term memory.