Depending on the kind of testing you're hoping to do you may want to look into https://github.com/mguenther/kafka-junit. For example, you're looking for some job level smoke tests that just answer the question "Is everything wired up correctly?" Personally, I like how this approach doesn't require you to open up the design for the sake of testing.
On Tue, May 24, 2022 at 8:34 AM Piotr Domagalski <pi...@domagalski.com> wrote: > > Hi, > > I'm wondering: what ithe recommended way to structure the job which one would > like to test later on with `MiniCluster`. > > I've looked at the flink-training repository examples [1] and they tend to > expose the main job as a class that accepts a `SourceFunction` and a > `SinkFunction`, which make sense. But then, my job is normally constructed > with `KafkaSource` which is then passed to `env.fromSource(...`. > > Is there any recommended way of handling these discrepancies, ie. having to > use `env.addSource(sourceFunction)` vs `env.fromSource(source)`? > > [1] > https://github.com/apache/flink-training/blob/05791e55ad7ff0358b5c57ea8f40eada4a1f626a/ride-cleansing/src/test/java/org/apache/flink/training/exercises/ridecleansing/RideCleansingIntegrationTest.java#L61 > > -- > Piotr Domagalski -- Cheers, Aeden GitHub: https://github.com/aedenj