Troubleshoot Import
Resolve rejected CSV files, mapping and preview errors, failed runs, partial completion, and history cleanup problems.
Use the file state, mapping, status, counts, and visible errors to choose the narrowest recovery. Do not test another organization’s Import identifier or broaden destination access to make a run pass; permission and destination-access rules remain under verification.
The CSV file is rejected
Import accepts one .csv file with a maximum size of 50 MB (50,000,000 bytes). A rejected file appears as failed and blocks Next.
- Remove the rejected item.
- Confirm the filename ends in
.csvand the exact file size is within the limit. - Upload only that file and wait for the ready state before continuing.
Mapping cannot continue
- Map every required destination field before selecting Next.
- Check that the first CSV row contains the intended headers.
- If a header mapped to the wrong field, return to Map columns and correct it before starting the run.
- Recheck the final mapping on Review and import; the review page is the last confirmation before rows can be written.
Preview or start fails
| Message | Recovery |
|---|---|
| Could not preview the import | Keep the draft and use Previous to verify the target, context, file, and mapping. Preserve the organization, Import name, time, and visible message. Do not expose or guess the logged reason; contact support if the generic message persists. |
| Could not start the import | Keep the draft and use Previous to verify the target, context, file, mapping, and Import name. Do not repeatedly start it; preserve the visible details and contact support if the generic message persists. |
| The preview says rows are blocked by validation errors | Correct those rows in the CSV and re-upload it. Do not confirm the preview while blocked rows remain. |
A run stays Processing
The detail page refreshes while a run is Processing. Wait for a final state before starting a duplicate import. If it remains active beyond the expected window, record the organization, Import name, start time, filename, and visible status in an approved private support channel. Do not attach customer data unless that channel and process explicitly allow it.
A run is Completed with errors
Completed means the processor finished; it does not guarantee that every row succeeded. Review Failed records and Error details, then reconcile the destination. The Success rate can count inserts, updates, matched rows that were unchanged, and skips as successful processing.
Correct failed source rows and start a new import for the corrected data. Do not assume rerunning the whole file is harmless when the conflict strategy can update existing records.
A run is Failed
A run-level failure shows Import failed and the visible error. There is no generic user-facing retry for the same failed Import record, but the background worker can make up to three total attempts. The page can therefore show Failed before automatic attempts are exhausted.
- Preserve the visible error and run details.
- Do not start a replacement import yet. Contact support to confirm that no background attempt remains.
- Reconcile destination records and any other persisted changes across attempts.
- Correct the input, mapping, or destination setup, then start a replacement import and validate its result independently.
Delete did not undo the import
This is expected. Delete removes the Import record and its history only. The confirmation explicitly says that entries created from the import will not be deleted. Use the destination product’s approved correction or deletion workflow for created destination rows.