Prerequisites
Before you begin, make sure you have:
- A Lakebase Autoscaling project accessible from both the Databricks UI and CLI. Commands identify it by project name, not by host.
- The Databricks CLI, authenticated with a profile for that workspace.
psqlavailable on yourPATHfor the SQL steps.- Podman or Docker with the compose plugin — required only for the optional Grafana dashboards.
Lakebase CDF is enabled later, in Step 4, rather than beforehand.
It operates on the lakets_cdf schema and its shadow table, both of which LakeTS
provisions for you: the schema during install, and the shadow when
enable_sync runs during setup.
Authentication
No static passwords are required. The jobs authenticate with machine-to-machine OAuth
automatically; for psql, you generate a short-lived token (covered in the next
step). Grafana is the exception — it cannot refresh tokens, so its setup
page describes the alternative.