Fix a failed or stuck import

Read the import outcome notifications, use Re-run import, and diagnose the usual suspects: dead feed URLs, changed columns, junk filters and rate limits.

Start from the notification

Every import run reports its outcome to the bell under Imports & syncs: "Import completed", "Import finished with warnings", "Import finished with errors" or "Import failed" (price & stock syncs report the same way). Failure and error rows carry a one-click Re-run import button — for transient problems (a feed URL that was briefly down, a marketplace hiccup) that's usually all it takes.

Common causes by source

• File / URL feeds — the feed URL stopped responding, its credentials changed (re-enter them on the Feed description step), or the file's structure changed so mapped columns no longer exist. Re-open the import from the card's Edit Import Settings and re-fetch: the structure detection will show what changed.
• Rows missing, not failing — check your junk filters: skipped rows aren't errors, they're filtered. Quick Scan shows how many rows your rules drop.
• Shopify — "Shopify import failed" retries automatically; if it persists, the store connection usually needs re-approval (see Fix a disconnected marketplace).
• Etsy — "Etsy import paused" is not an error: Etsy's request limit was reached, everything read so far is saved, and running the import again picks up where it stopped.
• Amazon / eBay pulls — pulls run in the background (Amazon exports a report first), so "nothing happened yet" a minute in is normal; the completion notification tells you when rows have landed.

Is it stuck, or still working?

The marketplace card tracks the run live — "Started products Import", then "Completed products import". An AI run has its own live progress and log on the Processing Job step, including a Stop Import button; stopped items are marked failed so you can re-run them. If a run shows no progress for a long time and no completion notification arrives, re-run it — imports are safe to repeat, existing rows are updated rather than duplicated.

Still stuck?

Email [email protected] with the notification text and roughly when the run started — that's enough for us to find the job on our side.