The test data is generated in a flink program running in a separate jvm. The generated data is then written to a Kafka topic from which my programs reads the data ...
|
Makes sense. The generation process seems to be inherently faster than the consumption process (Flink program). Without backpressure, these two will run out of sync, and Kafka does not do any backpressure (by design). On Thu, Sep 24, 2015 at 4:51 PM, Rico Bergmann <[hidden email]> wrote:
|
Free forum by Nabble | Edit this page |