/v1/videos 409 job_failed vs job_not_completed in retry loops

Both errors are HTTP 409 on GET /v1/videos/{id}/content, so a retry loop must read the code. job_not_completed says the job is still running: it advertises retryable: true and a poll_status next action. job_failed says the job ended in failure: retryable: false and next_action: "inspect_events".

The two 409s

The difference is the whole point of having two codes.

Why two codes

The docs say job_failed is intentionally not job_not_completed. Otherwise a client would poll forever for content that will never exist.

A retry-safe download

This script fetches the content and prints the error code if the answer is a 409. The code field is read defensively because the exact error body shape is not spelled out in the docs.

import asyncio, os
import httpx

async def main():
    headers = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
    job_id = os.environ["JOB_ID"]
    async with httpx.AsyncClient(base_url="https://api.sume.com", headers=headers, follow_redirects=True) as c:
        r = await c.get(f"/v1/videos/{job_id}/content", params={"index": 0})
        if r.status_code == 409:
            body = r.json()
            err = body.get("error") if isinstance(body.get("error"), dict) else body
            print("409:", err.get("code"))
        else:
            r.raise_for_status()
            open("out.mp4", "wb").write(r.content)

asyncio.run(main())

Poll first

Better still, poll GET /v1/videos/{id} until the status is terminal and only then ask for content. On failed, read the poll error string, which is the public remap of the failure.