Good Morning | Afternoon Team, my flink (running on KDA) app is reading data from a Kafka source, there are 12 partitions on the kinesis topic that i am reading and i have 30 KPU on KDA. When flink read from kafka only 12 KPU is conumming data and others don’t, can someone please shade some light on me that how can i use all of my KPU to consume data ?

No that’s not possible. A partition is meant to be consumed by one consumer. This isn’t a limitation in flink. Why not increase the partition number to match your flink parallelism

Your source parallelism cannot be more than the number of partitions. This is the expected behavior. You can limit your source operator to have a parallelism of 12 and increase the parallelism for other downstream operators

thank you so much for both of you for suggesting workaround, i will check those out flink