# Pipeline control panel # Pipeline control panel Data Runs the document-processing pipeline on request: resolves which documents actually exist for a set of records, checks what has already been processed, dispatches only the remainder, replays failures, and confirms the results reached their destination. - **Used when** — When a batch of records needs its documents processed, when a previous run left failures behind, or when someone needs to know whether a given account's documents made it through. - **Loads first** - Its own dispatch workflow, followed step by step rather than improvised. - The integration guide for the source system holding the document rows. - The read patterns for the store that records what has already been processed. - **Returns** - The source environment, the processing target, and the run identifier. - Input row counts, unique record counts, and how many fall outside supported handling. - Document-metadata counts, storage verification counts, and any location that could not be read. - The already-processed count and the count remaining to dispatch. - Dry-run and live dispatch counts, the queues targeted, and the manifest written. - Verification of the processed results, and a summary of any follow-up requests raised. ## Others in Data - Counterparty document expectations - Data-lake assistant - Dataset fulfilment - Knowledge search - Metric reconciliation - Model curator - Property data exploration - Public record ingestion - Question answering - Report catalog - Search to production