apache kafka – Does the union operator ensure same rate of consumption from all datastreams?

Does the Apache Flink’s union operator ensure that the rate of consumption from the data streams is same?

If the data streams are kafka topics, can a scenario happen when only one of the topic is being read through and no messages are being read from the another?

If so how can I detect and prevent it?

Put other ways I want to make sure that the watermarks (event-time) of the two data-streams in the union, is the same or within a window and that watermark is only X hours behind the current epoch time.