Production Task Ops: Query, List, Cancel, Delete
- →Operate on the real five-value status enum and route each status to the correct action
- →Use the single DELETE endpoint correctly and predict which task states error out
- →List and paginate tasks with the query endpoint's filters, and keep your own records beyond the 7-day window
- →Budget concurrency as a semaphore against the CONN limit instead of retrying blind
Everything in production H3 work keys off five status strings: queued, running, succeeded, failed, cancelled. Three are terminal — succeeded, failed, cancelled — and the documented polling interval is 10 seconds.
Cleanup is a single endpoint: DELETE /v2/video_generation/{task_id}. The task id is a path parameter, the HTTP method is DELETE, and there is no request body. You do not choose between "cancel" and "delete" — the server decides from the task's current status:
Unlock “Production Task Ops: Query, List, Cancel, Delete”
The 4-lesson Foundation module is free and always will be — you have already read the part most guides get wrong. The paid lessons go deeper, with labs, knowledge checks, and every technical claim verified against live documentation.
Already bought it? Sign in with the email you used at checkout.