Mailgun

What Alphorn adds on top of Mailgun

Wiring Mailguninto your application directly is the easy part. What usually grows into a maintenance problem is everything around it — retries when the provider is having a bad day, backoff so you don't get rate-limited, fan-out to additional destinations when a rule fires, and an audit trail for compliance or debugging. Alphorn owns all of that on your behalf, so the integration code you write stays at one HTTP call forever.

Every notification is placed on a durable queue before it touches the Mailgun API. If delivery fails, Alphorn retries with exponential backoff up to five times and records the failure in the delivery log. You can trigger a secondary notification when deliveries to a primary channel fail — for example, page an on-call engineer when a daily summary to a chat room stops going through. Filtering rules based on priority, tags, or JSON path expressions decide at runtime whether a particular payload reaches Mailgun or gets routed somewhere else entirely.

The same rules apply whether you're on the managed edition or running Alphorn yourself. The self-hosted edition is AGPL-3.0-or-later and ships as a single container, so you can deploy the exact same delivery pipeline on your own infrastructure — useful for regulated environments where outbound notification data cannot leave a private network unencrypted.