Full Refresh bei verwalteten Datenbank-Konnektoren wie SQL Server verzögert das Tabellen-Refresh bis nach dem Snapshot, hält die Ziel-Streaming-Tabellen währenddessen verfügbar und wendet gesammelte Änderungen gebündelt an, was PENDING_RESET- und Timeout-Fehler reduziert.
Full refresh operations for managed database connectors (like SQL Server) now reduce downtime by delaying the table refresh until after the snapshot is completed. During the snapshot process, destination streaming tables remain available with their existing data. New inserts, deletes, and updates are accumulated but not immediately applied. After the snapshot completes, the full refresh and other accumulated changes are applied in a single update, minimizing disruption to data availability. This improvement helps reduce PENDING_RESET and timeout errors during full refresh operations. See Full refresh behavior (CDC).