Retry a failed delivery on a customer-owned webhook
post
/v1/me/webhooks/{webhookId}/deliveries/{deliveryId}/retryFlips the row back to 'pending' with attempt=0 and next_retry_at=now() so the WebhookDeliveryWorker re-attempts it on the next tick. Only acts on rows currently in 'failed' — pending will be retried by the worker on its own schedule, success doesn't need a retry. Mirrors the admin retry endpoint but is scoped to the caller's org and limited to deliveries that actually belong to the named webhook (cross-tenant existence is never leaked).