Resend delivery
Sends a delivery again now, with the same body and the same event id, and returns the new delivery with the answer of your server. It makes one attempt and does not count as a failure. It is possible only while the body is kept, that is 7 days after the outcome (409 webhook_delivery_expired after that), and only when the endpoint is active (409 webhook_endpoint_inactive). Your key needs the read scope of the event family. The number of resends per hour for each organization is limited (429 rate_limited). With a test key nothing is sent and simulated is true. Required scope: webhooks:manage. With a test key (vx_test_) it is simulated: nothing real happens and nothing is billed. Requires an Idempotency-Key header. Retriable: when the operation itself fails with a 429, a 5xx or a temporary 409, the key is released and a retry with the same key runs the operation again; with no response at all, a retry with the same key returns the first outcome or runs the operation again. A 429 rate_limited from the request limit and 409 request_in_progress come before the operation and leave the key as it is: retry with the same key.
Authentication
Path parameters
Id of the delivery, as returned by GET /webhook-endpoints/{endpoint_id}/deliveries.
Headers
API version for this request; defaults to the version the key is pinned to.
A unique value you generate for each new operation (16 to 64 letters, digits, hyphen or underscore; a UUID works). Repeating the same request with the same key within 24 hours returns the first response and does nothing again, when that response was a success or a 4xx error the same request would always get again (not 429, and not the temporary 409 codes request_in_progress, caller_has_active_call and caller_request_in_progress: for those, and for a 5xx, see the note on retrying in the operation); the same key with a different request is refused with idempotency_key_reused.
Response
Always webhook_delivery.
Id of the delivery: use it in GET /webhook-deliveries/{delivery_id}.
Id of the event. It is the same on every attempt and on a resend, and it is the VoxAI-Event-Id header: use it to ignore an event you already processed.
The type of the event, for example call.ended; webhook.test for a test event.
pending: waiting for the next attempt. delivered: your server answered with a 2xx status. failed: every attempt failed and no more will be made. discarded: the feature was switched off before it was sent.
When the next attempt is due, ISO 8601 UTC. null unless status is pending, and always null for a test or a resend.
HTTP status of the last answer of your server; null if it did not answer.
A short safe code for the last failure, null if the last attempt succeeded: http_<status> (for example http_503), timeout, network, dns, address_not_allowed, port_not_allowed, internal_error, feature_not_available, expired, attempt_lost, error. Never text from your server.
How long the last attempt took, in milliseconds; null if there was none.
A test event or a resend: it counts neither as a failure nor towards the automatic disabling (a success resets the counter).
Id of the delivery this one is a resend of; null if it is not a resend.
true if the body carries the transcript of the call.
When the delivery reached its outcome, ISO 8601 UTC; null while it is pending.
true with a test key: the request was validated and NOT sent.
The exact JSON text that was signed and sent, while it is still kept (7 days after the outcome). null after that, and for a key without the read scope of the event family.
At most the first 1 KiB of your server’s answer, if it sent any. null for a key without the read scope of the event family.