C#/.NET Developer to Fix Duplicate and Missing Rows in Azure SQL Sync

We need an experienced C#/.NET developer to fix a reproducible bug in an existing data-sync service running on Azure App Service. The .NET 8 service incrementally synchronizes one customer table into Azure SQL. Rows sharing the same `ModifiedUtc` timestamp can occasionally be skipped. Retrying a partially completed batch can also create duplicate records. The application already builds, and we will provide: - Repository access - Staging environment - Sanitized test data - Reproduction steps and logs - Existing deployment pipeline Scope: 1. Diagnose the current watermark, checkpoint, and upsert logic. 2. Prevent records with identical timestamps from being skipped. 3. Make retries idempotent using the existing `ExternalCustomerId`. 4. Add or adjust a unique constraint if appropriate. 5. Add focused automated tests. 6. Deploy the fix to the existing staging environment. 7. Provide a brief explanation of the root cause and changes. Acceptance criteria: - A 30-record test batch produces exactly 30 unique records. - Reprocessing the same batch creates no duplicates. - New records sharing the previous watermark are not skipped. - Existing records are updated rather than duplicated. - The sync recovers correctly after a simulated partial failure. - The checkpoint advances only after successful processing. - Existing tests continue to pass. This project covers one synchronization process and one table. It does not include a pipeline redesign or new infrastructure. Required experience: - C# and .NET 6–8 - SQL Server or Azure SQL - Azure App Service - T-SQL and incremental data processing - Idempotent upserts, checkpoints, and retry handling - Debugging existing production code When applying, briefly describe a similar incremental-load or duplicate-record bug you fixed and how you verified that no data was lost. A successful fix may lead to additional Azure and data-engineering work.