Make webhook logs last 3 days: keep Sume job ids and events yourself
Make keeps webhook logs for 3 days on standard plans and 30 on Enterprise. Store each Sume job id so you can audit a delivery after the log is gone.

Make keeps webhook logs for 3 days on standard plans and 30 days on Enterprise (Make webhooks, read 2026-10-02). A Sume video can be requested on Monday and questioned the next week, so write the Sume job_id to a table of your own and use Sume's job events as the audit trail.
Two Make settings that affect what you can look up
If you turn on the confidentiality setting, Make will show that a run happened but not what Sume sent. That is useful for privacy and unhelpful for debugging.
| Item | What the page says |
|---|---|
| Webhook logs, standard plans | 3 days |
| Webhook logs, Enterprise | 30 days |
| Keep data confidential | Payload data is not kept in execution logs; the run is still recorded |
| Store incomplete executions | Failed runs are kept for retry and count toward storage limits |
What Sume already records
Each Sume job keeps a public event timeline at GET /v1/jobs/{job_id}/events: created, queued, started, completed or failed, and webhook.delivery. Public events do not expose raw provider task ids or URLs. The event list is a pull snapshot, not a stream, so read it when you need it.
The job's webhook_delivery state on the receipt tells you whether delivery succeeded, and a Redeliver call re-sends the terminal event if the Make side lost it.
A minimal ledger
- At submit, save the
job_id, the scenario run, and the request time in a Data Store or your own database. - In the webhook scenario, update that row with the terminal status and result URL.
- Treat Make's log as short-lived debugging help, not as the record.
Honest limit
Sume does not hold your Make execution history, and Make does not hold the Sume job. Only the shared id links them. If you did not save the id, you can still list jobs with GET /v1/jobs, which returns the jobs your API key's member created, but matching them back to scenario runs by time is guesswork.
Sources
Related posts
More in Integrations
- Cline MCP remote not connecting: set type streamableHttp for Sume
Cline treats a remote server with no type as legacy SSE. Sume's hosted MCP is streamable HTTP and answers GET /mcp with 405, so set type to streamableHttp.
- Mistral connectors: confirm Sume's paid tools before they run
Add Sume as a Mistral custom MCP connector, then use tool_configuration include and requires_confirmation to keep paid generation behind a human check.
- n8n 3.0 binary data filesystem mode: keep the Sume result URL
n8n 3.0 stops supporting in-memory binary data and renames the folder to ~/.n8n/storage. Pass Sume result URLs instead of downloading big videos.
- n8n 3.0 Chat Trigger WebSocket frames: replying with a Sume job
n8n 3.0 sends every Chat Trigger WebSocket frame as JSON with a type field. How a chat reply should hand a Sume video job to the user.
Written by Sume