> ## Documentation Index
> Fetch the complete documentation index at: https://docs.spark360.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Retry Behavior

> Delivery status transitions, retry schedule, and backoff behavior.

The delivery worker processes queued rows and applies status-based retry logic without blocking core lead workflows.

## Processing states

* `pending` -> waiting for first attempt
* `failed` -> retry scheduled
* `delivered` -> terminal success
* `abandoned` -> terminal failure after retry policy or immediate 4xx

## Response handling

| HTTP Response    | Action                                                |
| ---------------- | ----------------------------------------------------- |
| `2xx`            | mark row `delivered`                                  |
| `4xx`            | mark row `abandoned` immediately                      |
| `5xx` or timeout | mark row `failed`, increment attempts, schedule retry |

## Retry schedule

| Attempt | Delay after previous |
| ------- | -------------------- |
| 1       | Immediate            |
| 2       | 1 minute             |
| 3       | 10 minutes           |
| 4       | 1 hour               |
| 5       | 4 hours -> abandoned |

<Warning>
  Worker row selection and guarded updates must both accept `pending` and `failed` statuses so successful retries can transition failed rows to delivered.
</Warning>