How to efficiently use Kafka in place of Sidekiq for background Job in Rails?

Question

Currently, I am to using sidekiq for background jobs in rails app. Now I want to use Kafka for the same using consumer-producer mechanism. For that I am using ruby-kafka and producing messages asynchronously.

My approach: In the perform method of MyNewJob < ApplicationJob Job I am calling produce method of Kafka and when consumer receives the message, I execute the actual piece of code(which was previously executing inside perform method) inside a new class method of MyNewJob.

It is working as expected. I want to confirm, is it the right approach or not? If not then what should be the right approach?

Pseudo Code:

I will call my job using any of MyNewJob.perform_now or MyNewJob.perform_later etc.

my_new_job.rb

class MyNewJob < ApplicationJob

  queue_as TestMixin::QUEUE_NAME

  rescue_from(App::ServerError) do
    retry_job wait: Rails.application.config.retry_interval
  end

  def perform(paylaod)
    #producing message asynchronously
    producer = kafka.async_producer
    producer.produce(payload, topic: "topic_name")
  end

  def self.run_job(payload)
    #code to executed previously written inside perform method above.
  end

end

consumer.rb (pseudo code)

  #client consumer topic subscribe logic and then start consuming.
  consumer.each do |message|
    MyNewJob.run_job(message.payload)
  end

I am also thinking of another approach where I skip bringing ApplicationJob class into existence and directly produce message in Kafka instead of calling Job.

Please confirm.

Edit 1: I am using ruby-kafka with asynchronous producer (kafka.async_producer).