How do I set retries without trashing my quota or the source API?

Default retry settings are often 'every minute, 5x'. Too aggressive for some APIs, too weak for others. Exponential backoff plus dead-letter is the robust setup.

Try this first

1. 1Per error type: 5xx and timeouts are retryable, 4xx (except 429) is not (input error, retry won't help).
2. 2Set exponential backoff: 30s, 1m, 5m, 30m, then dead-letter. That gives transient outages room without hammering the source API.
3. 3On 429 (rate limit): respect the Retry-After header, not your own schedule. That is what the source declares safe.
4. 4After last retry: write the payload to a dead-letter queue (DB table or Make Data Store or n8n table) plus error context. Don't let it disappear.
5. 5Daily review on DLQ: manually re-submit or permanently reject. A DLQ no one looks at is not a DLQ.

None of the above fits?

Describe your situation below. We pass your input plus the steps you already saw to our AI and return tailored next-step advice. If it's too risky to DIY, we'll say so.