How to Integrate a Kinesis Data Analytics Flink Application with a Self-Managed Kafka Cluster Running on Amazon EC2 (not Amazon MSK)?

It's fairly similar to reading from Amazon MSK. Consuming data from Kafka (or Amazon MSK) with Kinesis Data Analytics (KDA) basically means establishing network connectivity between brokers and task managers. With KDA you can enable VPC support so that the ENIs of the task managers are exposed in the subnets you have specified. If you then configure security groups of the ENIs correctly, your Flink app should be able to connect to the Kafka brokers. See the following references:

• https://docs.aws.amazon.com/kinesisanalytics/latest/java/vpc.html
• https://docs.aws.amazon.com/kinesisanalytics/latest/java/example-msk.html