Change history and pipeline replication

@cap-js/change-tracking records who changed what on entities you annotate for end-user auditing and UI “change history” features. cds-data-pipeline fills tables from remote sources and records per-run statistics on Pipelines / PipelineRuns (created, updated, deleted counts, watermarks, errors).

These concerns are orthogonal:

• Pipeline runs attribute bulk sync work to a run and a pipeline name, not to interactive user edits on individual fields.
• Change tracking attributes row-level changes to business users when they edit data through your app.

After a replicate run, rows in the target table match the remote slice you defined (consumption view, delta mode). Local enrichment columns (fields not overwritten by the pipeline’s UPSERT strategy) can still be edited by users; keep pipeline mapping and delta configuration aligned with what you want to preserve vs overwrite.

For pipeline observability, use the Management Service and Features → Observability rather than treating PipelineRuns statistics as a substitute for @cap-js/change-tracking.

See also

• Consumption views — local shape and what the pipeline writes.
• Built-in replicate → What happens at runtime — READ / MAP / WRITE and tracker updates.