In Atlas, the core concept of representing Flink applications, Kafka topics, HBase
tables, and so on, is called an entity. You need to understand the relation and definition
for entities in a Flink setup to enhance the metadata collection.
When submitting updates to Atlas, a Flink application describes itself and the entities it uses
as sources and sinks. Atlas creates and updates the corresponding entities, and creates
lineage from the collected and already available entities. Internally, the communication
between the Flink client and the Atlas server is implemented using a Kafka topic. This
solution is referred to as Flink hook by the Atlas community.