Reconciliation of streamed data

For streamed data, the manifests that are sent by the source application are directly published in Data Ledger.

No calculations are finished by Data Ledger before you can view the application's manifest. Therefore, when published, reconciliation is classified as unknown. To verify whether there is a data drift between the source application and Data Lake, you can query Compass from Data Ledger with a predefined Data Ledger query.

The match scenario happens only when the result of a Compass query is the same as the count published by the source application. Otherwise, reconciliation is classified as a mismatch.