my flink job now has 60+ output streams, each stream's implementation is simple。But, why back pressure happens in kafka source operator.
enter image description here my job graph
enter image description here kafka source back pressure
my flink job now has 60+ output streams, each stream's implementation is simple。But, why back pressure happens in kafka source operator.
enter image description here my job graph
enter image description here kafka source back pressure
Basically, it's physically impossible for your job to run faster than the throughput of the slowest sink.