Retries, timeouts and making your endpoint robust

What happens if your endpoint temporarily fails?

The Leafy Energy demo helps explain how a webhook dispatcher deals with timeouts and errors, and why you need a robust endpoint.

Typical retry strategy

• Leafy Energy waits for a response for up to about 5 seconds.
• On timeout or 5xx response, the event is added to a retry queue.
• The system retries several times using exponential backoff.

Best practices for your endpoint

• Process events asynchronously: store them and handle them in the background.
• Return a 2xx status as soon as possible.
• Log errors and monitor the health of your webhook service.

This article is perfect for discussing integration weak spots and introducing monitoring and alerting concepts.