How to clear a backlog after a connector outage
Use this procedure when the destination is healthy again and work remains queued.
What to do
- Count queued records, note their age, and identify dependencies or items that must stay in order.
- Resume a small batch and check the receiving service for errors or rate limits before increasing throughput.
- Reconcile completed, failed, and remaining counts. Review rejected entries for permanent mapping problems before any replay.
Reference: Microsoft Azure queue based load leveling pattern.
Related MyNew service: System Integration.