PATTERN Cited by 1 source

PR-based schema rollout

Pattern

Route all automated schema evolution changes through a pull-request workflow: the automation detects a schema change, regenerates all affected code and configuration artifacts, and opens a PR for review. The PR provides:

• Auditability — who, when, what changed
• Versioning — schema version tracked across the PR history
• Reviewability — engineers can inspect generated code before deployment
• Rollback path — revert the PR to undo the schema change

What the PR contains

In Pinterest's instance, the generated PR includes: - Updated Iceberg CDC + base table schemas - Regenerated Flink transformation code - Regenerated Spark writer code - Updated version metadata - Refreshed create and bootstrap queries

Why PR-based over direct-apply

Direct schema application (e.g., Iceberg ALTER TABLE) is fast but opaque. PR-based rollout trades speed for safety: the bounded SLA window is acceptable because schema changes are not consumed in real time. The PR also doubles as the deployment trigger — merged PRs drive CI/CD pipelines that deploy the generated code.

Seen in

• sources/2026-06-24-pinterest-automated-schema-evolution-in-pinterests-next-generation-db — canonical instance: all Pinterest CDC schema evolution changes flow through generated PRs