Sometimes a translation job in Elovate may finish with certain items showing errors or missing output. Instead of recreating the entire job, you can quickly retry failed translations directly from the job overview or item view.
1. Locate the failed items
Go to Apps → Translations → Jobs.
Find the translation job where failures occurred.
In the status column, you may see counts like 25/30 done — this means 5 items did not translate yet or not successfully.
Click on the job name to open it.
2. Filter for failed translations
Inside the job view:
Use the filter to show only translations with status “error”.
This will hide all successfully translated entities so you can focus on the problematic ones.
3. Retrying from the job view
Select the failed items you want to retry.
Click Regenerate.
The system will send those items back to the translation provider without affecting already completed translations.
4. Retrying from the item view
If you want to retry a single failed translation:
Open the specific item from the job.
Make any adjustments needed to the source text, target language, or prompt (for AI translations).
Click Regenerate translation.
5. Using the synchronisation log for troubleshooting
Go to Apps → Synchronisation Logs.
Every translation error is recorded here.
You can click directly on the affected entities from the translation job to inspect the error details.
This allows you to pinpoint the cause of the error (e.g., empty attribute, invalid language code, missing configuration) before retrying.
6. Common reasons for failures
GPT token limit reached – Add more credits and retry.
Invalid language pair – Adjust source/target to supported languages.
API timeout – Simply retry; sometimes provider delays cause failures.
Empty source attribute – Add source text before retrying.
7. Best practices for retrying
Batch retry to save time instead of retrying one-by-one.
Check your provider status before retrying to avoid repeated failures.
For ChatGPT, refine your prompt if results were inaccurate before.
Always review the synchronisation log before retrying — it often gives the exact fix needed.