Concepts

The small vocabulary cds-data-pipeline is built on. These pages cover the pieces of addPipeline(...), what the plugin infers from your config, and the event namespace every pipeline exposes.

• Terminology — Pipeline, source, target, mode, delta strategy, tracker, event namespace. The primitives every pipeline registration composes.

• Inference rules — How addPipeline(...) derives pipeline behavior from the config shape — read shape, inferred defaults — and which config combinations are rejected at registration.

• Configuration overrides — Persisted runtime overrides layered on the coded baseline (schedule, mode, delta, tuning knobs, pause). Survive restarts; surfaced in the management API and Pipeline Console.

• Run housekeeping — Opt-in retention for PipelineRuns history: global defaults, per-pipeline overrides, scheduled age/count pruning.

• Consumption views — The idiomatic CAP pattern for shaping a replicate target — a projection on <remote.Entity> that declares target schema, column restriction, and rename mapping in one place.

• Change history and pipeline replication — How @cap-js/change-tracking relates to pipeline-filled tables, per-run PipelineRuns statistics, and user-edited local enrichment data.