Eventsourcing-helpers

Minor:
- Bump `confluent_kafka_helpers` verison to `0.7.5`
- Rename serialized message key `meta` to `Meta` (https://github.com/fyndiq/eventsourcing_helpers/pull/62)
- Use `dequeue` as data type in mock messagebus backend (https://github.com/fyndiq/eventsourcing_helpers/pull/61)

Features:
- `messagebus`: add mock backend (https://github.com/fyndiq/eventsourcing_helpers/pull/56)

Fixes:
- `messagebus.kafka`: define top level config keys to use when creating producer and consumer configs (https://github.com/fyndiq/eventsourcing_helpers/pull/60)
- `message`: keep keys with none values in `to_dict` method (https://github.com/fyndiq/eventsourcing_helpers/pull/59)

Minor:
- bump `confluent_kafka_helpers` verison to `0.7.4`
- make `cnamedtuple` an extra instead of a dependency (https://github.com/fyndiq/eventsourcing_helpers/pull/55)

* Use new version of confluent_kafka_helpers

* Reduce the default fetch request interval for kafka consumers. The purpose is to reduce network traffic and GCP costs. In time critical services (Inventory, Accounting) we must override the configuration for fetch.wait.max.ms`. 10 ms is a good value to start with for time critical services.
* Make Redis configurable from application
* Refactor the aggregate builder using the new repository interface
* Rename ESEntityBuilder to AggregateBuilder. In ES you build an aggregate, not an entity.

- Bump confluent_kafka_helpers to 0.7.0