Governance, performance, and capstone · Lesson 1 of 3
Unity Catalog access and privileges
Apply least-privilege grants to governed tables and pipeline identities.
Lesson
Unity Catalog provides a governed namespace and privilege model for data assets. Access decisions involve the principal, object, and required privilege; users and service principals may receive privileges directly or through groups. A production job should run with an identity that has only the grants needed to read sources and write targets. Separate ownership from routine use and review inherited access when troubleshooting.
Grant the narrowest scope that supports the workflow and avoid broad workspace-wide permissions as a shortcut. A notebook author's access does not prove that a job's service identity can access the same table. Practice task: list the minimum source-read and target-write permissions for a pipeline principal, then identify how to test the job using that identity.
Applying the pattern: Identify the job principal, inspect object ownership and grants, assign minimum privileges, and test both allowed and denied operations.
Checks and failure modes: A successful interactive query may be using a human identity with broader access than the production job.
Before production, test both a representative input and a boundary case. Confirm the expected row grain, permissions, recovery behavior, and observable output for the task. Record the configuration needed to reproduce the run and compare its result with the agreed data contract.
Practice task: Which identity should be checked when a scheduled pipeline gets an access error?
Visual guide
Authorize a pipeline operation
-
Step 01
Principal
Identify the job run identity.
-
Step 02
Source
Grant required read access.
-
Step 03
Target
Grant scoped write access.
-
Step 04
Verify
Test effective access in the job.
Key points
- Identify the job principal, inspect object ownership and grants, assign minimum privileges, and test both allowed and denied operations.
- A successful interactive query may be using a human identity with broader access than the production job.
- Re-run the included practice scenario after changing source data or transformation logic.
Check your understanding
Practice
Your learning progress