Meter Ingest Automatic Failover

0 min

For continuous and resilient production ingestion, Amberflo provides a high-availability endpoint: ingest.amberflo.io.

This endpoint is backed by a multi-regional deployment, ensuring that even if the primary ingestion region experiences issues, your meter events will still be received and processed through a failover path.

To take full advantage of this failover capability, we recommend implementing a retry policy in your API client. Retry guidelines:

Use an exponential backoff strategy

Ensure at least a couple of retry attempts are made after 20 seconds from the first failure

Retry specifically on the following HTTP status codes:

429 (Too Many Requests)

Any 5xx server error

If you use one of our SDKs, they have built-in logic for a suitable retry policy.

3.1.0 for Python

2.6.0 for TypeScript / JavaScript

2.1.0 for Go

2.1.0 for Java