Cleanup prometheus metrics after Rebalancing

[marcosschroh]

Describe the bug
After a Rebalancing the StreamEngine could contain metrics of partitions that where assigned to other StreamEngine instances which are running in different node. We need to make sure that we clean the old metrics.

To Reproduce
Start with x amount of instances, scale up/down and the StreamEngine contains old metrics.

Solution

Use the ConsumerRebalanceListener to listen for Rebalancing events. We have an open PR on aiokafka to be able to create a consumer with a ConsumerRebalanceListener.

[marcosschroh]

Closed by #104