Documentation Index
Fetch the complete documentation index at: https://tracefinance-docs-withdrawal-beneficiary-events.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
Overview
When your endpoint fails to acknowledge a delivery — non-2xx response, network error, or timeout — Trace records the failure and queues a retry. Every attempt is captured in an execution log you can inspect or replay manually.
What counts as success
Trace considers a delivery successful when your endpoint returns any2xx status code (200, 201, 202, 204, etc.). Anything else — 4xx, 5xx, network failure, or response timeout — is recorded as FAILED and triggers retry.
Retry behaviour
Failed deliveries are queued and re-attempted after a delay. Trace retries the same payload with the same headers and signature; from your side, a retry is indistinguishable from the original delivery — except thatX-Message-Id stays the same, letting you detect duplicates.
| Property | Behaviour |
|---|---|
| Triggers | Any non-2xx response, exception, or network failure |
| Backoff | Fixed delay between attempts |
| Max attempts | A small number per delivery (typically 2–5) |
| Identifier | X-Message-Id is constant across retries — use it for idempotency |
| Result | After the budget is exhausted, the execution log stays FAILED and no further automatic retries occur |
Retry counts depend on the deployed environment. Reach out to your Trace contact if you need the current limit for capacity planning.
Idempotency on your side
Always treat handlers as idempotent. The sameX-Message-Id may arrive multiple times if:
- Your endpoint timed out but actually processed the event.
- A previous delivery returned
5xxand is being retried. - You manually resend a delivery.
X-Message-Id somewhere durable (a database, cache, or queue) and short-circuit duplicates before doing any side-effects.
Inspect execution logs
Every delivery attempt is recorded with the request payload, response status, and retry count.Manual replay
If your endpoint was unavailable during the original retry window, you can replay any execution log on demand. Trace re-sends the same payload, headers, and signature.204 No Content if the resend is queued. The replay is recorded as a new attempt in the execution log history.
Related
- Verify signatures — duplicate-safe processing depends on a correctly identified message
- Subscribe to events — pause delivery instead of returning
5xxwhen you need a maintenance window - Resend a delivery — full reference for the manual replay endpoint