Good day Rahma Nali
During large‑scale migrations, the CRM deduplication engine often becomes the bottleneck. The supported procedure is to temporarily disable the record matching and deduplication logic so the batch ingestion can run without locking the index. Most CRM platforms provide a “bulk insert” or “disable duplicate detection during bulk operations” flag that you can enable for the import job. This way, the system skips deduplication checks while loading data, and you avoid the deadlocks. Once the migration is complete, you can re‑enable deduplication and let the engine reconcile records. It’s also smart to monitor database performance during the import, since contention can still occur if the batch size is too large. Splitting the ingestion into smaller chunks can help reduce latency. Afterward, run a post‑deduplication process to clean up any duplicates created during the bypass.
If this explanation helps you unblock the pipeline and reduce latency, please hit “accept answer” so we know it was useful!