
Stock Sync
Part of Dropshipping stock synchronisation
Monitoring failed inventory updates
An inventory feed can arrive on time while some rows fail to update.
An inventory feed can arrive on time while some rows fail to update. A useful monitor checks the store result, not just that a file downloaded or an API request was sent. In Shopify that means three checks: the error fields returned by each write, the inventory_levels/update webhook that confirms a change, and a scheduled comparison of feed and storefront quantities.
Record each run
Capture the source timestamp, arrival time, row count, mapped variants, unknown SKUs, rejected rows and completed updates. Compare these with the prior successful run. Write down two store thresholds, one for a fall in row count and one for a large single-variant quantity change, set from your normal run-to-run variation. Flag the run for review when either is crossed, even if the import reports success.
Shopify's inventory CSV guidance covers exporting current quantities, updating them in the CSV and importing the file to update quantities. Its inventorySetQuantities mutation uses compare-and-set behaviour and returns an error if the compare quantity does not match. Each response includes userErrors with a code, field and message; log all three against the SKU and location.
Since the 2026-04 API version, inventorySetQuantities requires an idempotency key. Log the key with each write, so a retried request can be matched to the original instead of being counted as a second change.
A sample failure record might read: run started 06:00, feed rows down from 8 to 7, one SKU unknown, one write returned a userError on its compare quantity, storefront still showing 3 available. Resolution: mapping corrected, write retried with the same key, storefront checked at 0, owner and time noted.
Key Metrics to Monitor in Inventory Sync
- Row Count Variance Threshold
- Set based on historical run-to-run variation
- Unknown SKUs Detected
- Flag if >0 – indicates mapping issues
- User Errors in API Response
- Log all codes, fields, messages for troubleshooting
- Idempotency Key Required
- Since 2026-04 API version – must be logged
Detect silent failures and escalate by sales risk
Subscribe to the inventory_levels/update webhook. Its payload carries the inventory_item_id, location_id, available quantity and updated_at time, so each write you sent can be matched to a confirmed change. A write with no matching webhook after the run is a silent failure, even when the import showed no error.
Alert at once, to a shared channel with a named owner, when a failed update leaves a high-selling or near-zero variant with stale positive stock. Record the exact SKUs and affected channels so the operator can pause sales or correct the mapping. Lower-risk failures can go into a daily summary. Distinguish a supplier-feed failure from a storefront-write failure; they need different owners.
After a repair, rerun a controlled update and verify the storefront state. Keep an audit trail of the failure and resolution rather than treating the next successful run as proof that no orders were affected.
Detecting Silent Failures vs. Successful Imports
- Import shows success but no webhook
- Silent failure – inventory not updated despite successful API response.
- Webhook received
- Confirmed update – inventory level changed as expected.
- High-selling item with stale stock
- Immediate alert required – risk of overselling.
- Low-risk variant with minor error
- Include in daily summary, no urgent action needed.



