Apache Flink

:wave: I’m a newbie trying to figure out how the parallelism is supposed to work using the K8s operator. I’m working based on the <https://github.com/apache/flink-kubernetes-operator/blob/ad8b4abf7334c7f196bd5a43f43e6181c9466574/examples/basic.yaml|basic example>.

How is the `parallelism`  setting in the `job`-spec supposed to work?

If I change the replica count on the `taskManager` or change the `taskmanager.numberOfTaskSlots: "2"`  but keep the `job.parallelism` setting as is, the parallelism of the job (according to the UI) changes.
Changing the `job.parallelism`  does not seem to have an effect.

I tried to read the <https://nightlies.apache.org/flink/flink-kubernetes-operator-docs-release-1.4/docs/custom-resource/overview/|documentation for the CRD> but it did not clarify the issue.

Am I misunderstanding something?

Nevermind, I figured it out by reading the documentation closer.

The problem was that I had configured the `replicas` for the task manager, and the <https://nightlies.apache.org/flink/flink-kubernetes-operator-docs-main/docs/custom-resource/reference/#taskmanagerspec|documentation> says that:
```replicas 	java.lang.Integer 	Number of TaskManager replicas. If defined, takes precedence over parallelism```
Once I removed the replicas configuration, the `job.parallelism` worked as expected!