Configure Metabase, LogReader, and UserDatabase components for one pipeline

Orchestrates Metabase, LogReader, and prepared-user-database configuration for one pipeline in a single call using the same connectionId-keyed request shape for both REPLICATION and ADVANCED_REPLICATION projects.

Features:

• Accepts three independent, optional arrays (metabases, logReaders, userDatabases); an omitted array leaves that component family untouched.
• Every entry is upserted (get-then-create-or-update) independently by its own connectionId and processed in parallel.
• Every submitted connectionId is pre-flight validated against the pipeline's actual connection membership; a non-member entry is marked FAILED before its processor is dispatched.
• For REPLICATION, each entry's connectionId is matched against the pipeline's actual source/target connection and routed through the exact same orchestration path as the dedicated Continuous composite endpoint — identical behavior, same request shape at this API boundary. Continuous supports at most one LogReader entry (bound to the source connection); additional entries are reported as FAILED (DI-ADAPTER-LGR-0006).
• For ADVANCED_REPLICATION, entries route through the DSL-specific processors with connection-type-aware parameter defaulting.
• A component failure is isolated and never blocks the others — the endpoint always returns 201 with a per-entry result (CREATED/UPDATED/FAILED plus errorCode when failed).

Use Cases:

• Configure every component a pipeline needs in one call after Full Validation.
• Re-run configuration idempotently; existing components are updated, not duplicated.

Authorization: Requires replication-pipeline::create or replication-pipeline::manage.