Skip to main content
POST
Enqueues up to 100 matching deliveries (oldest first) for asynchronous re-delivery and returns 202 Accepted — the request itself never performs outbound POSTs. The retry sweep re-delivers them at its own paced rate. Provide at least one filter. Each re-delivery keeps its original event_id, so receivers dedupe.

Body Parameters

Provide at least one:
string
Re-deliver all deliveries of one logical event (across endpoints).
string
Re-deliver deliveries for one endpoint.
string
Re-deliver deliveries in a state: pending, delivered, failed, dead.
string
Lower bound on created_at (ISO-8601, timezone-aware).
string
Upper bound on created_at (ISO-8601, timezone-aware).
A delivery whose endpoint is inactive is skipped (not an error). Re-running the same filter is idempotent: a delivery that already has a live scheduled retry is reported as skipped, never double-fired.

Examples

truncated is true when more than 100 deliveries matched. last_created_at is the created_at of the last delivery this call selected — pass it as the next call’s date_from to advance past the page you just enqueued.
Bulk replay is capped at 100 per call and rate-limited per account. To back-fill a large range, walk it with date_from = last_created_at while truncated is true. Repeating the same filter is safe but does not advance: selection is oldest-first and a requeued delivery keeps its created_at, so the same page is re-selected and reported as skipped until the sweep delivers it.

Error Codes