Retry mechanism for failed webhooks

[eelcoj]

Just a manual retry from the admin UI could be nice.

[primeautomation-dev]

Manual retry from the admin UI makes sense 👍

One thing we learned in production: UI retry should not directly re-execute the handler.
It should enqueue a retry with the same idempotency key and go through the normal delivery flow.

That way:

• manual retries stay safe under concurrency
• crashes mid-execution don’t cause duplicate side effects
• auditability stays intact (who retried, when, why)

Happy to help sketch this if useful.

[eelcoj]

Thanks for the offer @primeautomation-dev (even more so as it super early days! ❤️ ). Enqueuing sounds good, but currently Fuik has/expects no ActiveJob dependency. Tracking too, sounds good, but maybe a bit too much for now—it needs a fair amount of more changes to the data model.

Sketch/ideas are helpful. Thanks!