Just wondering how a REST API can be on the consumer end? The consumer must make a call to the MPS to get a response. Are we setting up the application in a way where periodic REST calls are made to MPS and then MPS reads from the Kafka queue? This seems like a pull model than a push model.
How is Poison Pill a problem , because consumer if fails to process after retrial can move the message out and start processing other messages. If autocommit is on , the even before processing it commits the offset to the broker and even if its failed its not picked back again. in case of autocommit off , if it fails is moved out to DLQ , why will the consumer go back to fetch the same batch again , i have mostly used KAFKA and it doesn't do that ever...
Great article!
Just wondering how a REST API can be on the consumer end? The consumer must make a call to the MPS to get a response. Are we setting up the application in a way where periodic REST calls are made to MPS and then MPS reads from the Kafka queue? This seems like a pull model than a push model.
It seems other way arround. where MPS is calling the consumer Rest endpoint and sending the msg as api payload. its push method from MPS to Consumer.
this can be hanndled by a topic with its consumer end point.
How is Poison Pill a problem , because consumer if fails to process after retrial can move the message out and start processing other messages. If autocommit is on , the even before processing it commits the offset to the broker and even if its failed its not picked back again. in case of autocommit off , if it fails is moved out to DLQ , why will the consumer go back to fetch the same batch again , i have mostly used KAFKA and it doesn't do that ever...
Great article.
For a new perspective on technology, read this:
https://swiftenterprises.substack.com/p/computational-independence
Greeting everyone my name is igwe chigbo Tboy beautiful
Not to much beautiful, tbh