Rate limiting and retry policy for streaming destinations built with Destination SDK

Partner-built destinations can return various errors and have different rate limiting policies. This page explains how Experience Platform handles different types of errors returned by streaming destinations.

When configuring a destination using Destination SDK, you can select between two aggregation types - best effort aggregation and configurable aggregation. Depending on the aggregation type that you select, read below how Experience Platform handles errors and rate limitations.

Best effort aggregation best-effort-aggregation

For any HTTP calls made to your destination that fail, Experience Platform attempts to make the call again one more time immediately after the first call. If the call still fails on the second attempt, Experience Platform drops the call and does not reattempt it a third time.

Configurable aggregation configurable-aggregation

In the case of destination platforms set up with configurable aggregation, Experience Platform distinguishes between the error type returned by your platform:

Retry approach described retry-approach

The Experience Platform approach for configurable aggregation is described below. This example assumes that Experience Platform sends data to a destination platform that starts returning 429 error codes if it receives more than 50k requests per minute:

Next steps next-steps

You now know how Experience Platform handles errors and rate limiting from destination platforms, depending on the aggregation policy you selected when you configured your streaming destination. Next, you can review the following documentation: