Automate Monasca documentation for configuring Kafka

Until the Monasca Kafka client fork is removed it is currently required
to run Kafka in compatibility mode. It is also necessary to disable
an optimisation in the Kafka brokers to clean up idle connections. This
is because the optimisation was added after the Monasca Kafka client was
forked, and the client hasn't been updated since. These settings are now
applied automatically when Monasca is enabled.

Change-Id: I6935f1fb29f4f731cf3c9a70a0adf4d5812ca55e

ansible/roles/kafka/templates/kafka.server.properties.j2

@@ -16,3 +16,7 @@ log.segment.bytes=1073741824
 log.retention.check.interval.ms=300000
 zookeeper.connect={{ kafka_zookeeper }}
 zookeeper.connection.timeout.ms=6000
+{% if enable_monasca | bool %}
+log.message.format.version=0.9.0.0
+connections.max.idle.ms=31540000000
+{% endif %}

doc/source/reference/logging-and-monitoring/monasca-guide.rst

@@ -46,14 +46,6 @@ following override in ``/etc/kolla/globals.yml``:
 
    monasca_install_type: "source"
 
-Until the Monasca Kafka client is upgraded it is currently required
-to run Kafka in compatibility mode. This can be achieved by adding some
-custom Kafka configuration:
-
-.. code-block:: console
-
-   echo "log.message.format.version=0.9.0.0" >> /etc/kolla/config/kafka.server.properties
-
 Stand-alone configuration (optional)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~