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

# Retry policies

> Understand delivery guarantees and configure retry behavior for webhooks.

The platform guarantees at-least-once delivery for webhook events. If your endpoint does not return a `2xx` response within the timeout window, the platform retries the delivery automatically using an exponential backoff schedule.

## Default retry schedule

Failed deliveries are retried up to 5 times over approximately 24 hours:

| Attempt   | Delay after previous failure |
| --------- | ---------------------------- |
| 1st retry | 5 minutes                    |
| 2nd retry | 30 minutes                   |
| 3rd retry | 2 hours                      |
| 4th retry | 8 hours                      |
| 5th retry | 24 hours                     |

After the 5th retry, the delivery is marked as permanently failed. A `sync.failed` event is triggered if the endpoint has repeated failures across multiple events.

<Info>
  Because the platform retries on failure, your endpoint may receive the same event more than once. Use the event `id` field to deduplicate and ensure idempotent processing.
</Info>

## View failed deliveries

To see a log of failed deliveries for an endpoint:

1. Go to **Integrations → Webhooks → \[Endpoint name]**.
2. Click the **Delivery log** tab.
3. Use the status filter to show only failed deliveries.

Each entry shows the event type, timestamp, response status code, and response body returned by your endpoint.

## Manual re-delivery

You can manually trigger re-delivery for any failed event:

1. Find the failed delivery in the delivery log.
2. Click **Re-deliver**.
3. The platform sends the original payload again immediately, independent of the retry schedule.

Use manual re-delivery after fixing an issue with your endpoint to recover events without waiting for the next automatic retry.

## Disable an endpoint

If an endpoint is consistently failing and you need time to fix it, temporarily disable it to pause delivery:

1. Open the endpoint settings.
2. Toggle **Active** to off.

Events that occur while an endpoint is disabled are not queued—they are dropped. Re-enable the endpoint once it is ready to receive traffic again.