kpl-callback-pool-* in KinesisProducer is not shutdown on destroy()

[ghost]

There are times when we need to restart the kinesis producer to get messages to send again, and we do this by calling KinesisProducer.destroy(), followed by new KinesisProducer(config). However, we notice that when doing this, the old kpl-callback-pool is still lingering around, so the more restarts we do, the more of these we have hanging around.

The threadpool should be shutdown when calling KinesisProducer.destroy()

[skidder]

This might be a duplicate of #93