Integrations: KSeF, payments, Allegro, ERP and couriers

Calls over the limit are refused and lost, because nobody retries them.

Events wait in the queue and leave at the pace the other side allows.

1. 1A limit on the other side: we wait as long as the other side asks says.
2. 2Events wait their turn in the queue.
3. 3The queue is empty again, no event was lost.

A call with no answer goes again blindly, and a second invoice or a second parcel appears.

A retry of the same event is recognised and nothing is created twice.

1. 1No answer in time: the same event is retried.
2. 2The retry never creates a second invoice, parcel or server.
3. 3Every order of the sale went through exactly once.

Everything sent during the outage was lost.

Events wait in the queue, retries go with a growing pause, and once the other side is back the queue empties.

1. 1A server error on the other side: retries with a growing pause.
2. 2A failure alert: we are the first to know.
3. 3The other side is back, the queue emptied in order.

From the switch-off, every call is refused until someone notices.

The integration moved to the new version before the date, so nothing stopped.

1. 1The provider announced the switch-off of the old API version.
2. 2Under maintenance the integration was adapted before the date.
3. 3Events go through without a break.