Sume video job failed: retry, new key, or switch the model?

A failed video job is final, and an Idempotency-Key replay returns the same failed job. Read the error, then retry with a new key or change models.

5 min readSume
All posts

When a Sume video job reaches failed, read its error field first. Sume's docs say an Idempotency-Key replay returns the original job, so sending the same key again gives you the same failed job, not a new attempt. To try again, send a new key. To avoid the same failure, change something about the request.

Decision table

After a failed video job (read 2026-10-03)
SignalAction
400 before a job existsFix the request; check GET /v1/videos/models limits
Job status failed with an errorRead the error; retry once with a new Idempotency-Key
Same model fails twiceSwitch to another model that lists your duration and resolution
Job status cancelledSomeone canceled it; do not auto-retry

Retry rules that protect your budget

The fallback and idempotency posts below have runnable Python for the retry and model-swap steps.

  • Cap retries at one per job.
  • Write your intent row before you submit so a retry reuses the same business record.
  • Never retry a 400; it will fail the same way.

Sources

Related posts

More in Developers

All Developers posts

Written by Sume