RabbitMq Ack, Nack

[conioX]

you can tell KrakenD to reinject the message (Nack) to retry the message later by any other worker. Notice that when working with Nack, if KrakenD is the only consumer and your backend fails to process the message continously, KrakenD will reinsert the message into the queue over and over

Hi I create consumer with RabbitMq and krakend, but if we create consumer with RabbitMq we can tell return ack, nack discard or nack requeue

How to tell krakend when we want nack discard or nack requeue?

[thedae]

Hi @conioX , if I'm following correctly, you're looking for a way to configure Krakend NACK strategy, right?

If that's the case, the current solution does not provide a way to customize the NACK behavior, you can only play with the auto_ack option in order to requeue a message on failure or always ACK regardless the outcome of the message consumption

[taniaelysia42]

yeah @thedae i try to implement https://www.krakend.io/docs/async/

so if my endpoint is return 200-299 it will ack? other will nack requeue?

because in rabbitmq as far i know we can NACK Discard and NACK Requeue not just requeue

[thedae]

so if my endpoint is return 200-299 it will ack? other will nack requeue?

That's correct

because in rabbitmq as far i know we can NACK Discard and NACK Requeue not just requeue

Yes, I understand this can be achieved in RabbitMq, the point here is that the current AMQP integration in Krakend does not allow to select between Nack discard/requeue,

[thedae]

In any case, thanks @taniaelysia42 and @conioX for the feedback, I'm closing this issue as it won't fix at this moment, but we're going to discuss it internally and evaluate if we want to introduce this feature in future releases