Does datahub requires Kafka service ?

Yes

Is there is a way to disable this ?

No, Kafka is a hard requirement for DataHub. In fact, DataHub was designed with Kafka as a central piece. There is on-going work to allow replacing Kafka with other alternatives such as Google pub-sub and removing dependency on confluent’s schema registry but I’m not aware of the progress on that.

What's the reason its a hard requirement for using Kafka,L?

Kafka specifically? I can only assume historical reasons and because DataHub came out of LinkedIn who are the original creators of Kafka. There was no reason to use anything other than Kafka. A messaging system like Kafka? To ensure high availability, resiliency between components and async processing which also enables scalability.

Initially we thought mce and mae consumer are using Kafka service. But now we figure out gms and frontend also requires Kafka And why these two images are using Kafka ?

Mce/mae consumer are Java applications that read Kafka messages, process them and may generate other messages. Gms is another Java application that generates messages for the aforementioned consumers to consume. Frontend I don’t know.

What msgs they are generating and who are the consumers ?

Among others, these: https://datahubproject.io/docs/what/mxe