GITBOOK-2: No subject

asynchronous-api-guidelines/kafka-asynchronous-guidelines/k_schemas.md

@@ -0,0 +1,31 @@
+# Schemas
+
+Kafka messages **SHOULD** use schemas (AVRO, JSON, Protobuf) registered in the Schema Registry to ensure compatibility between producers/consumers.
+
+If so, always refer to the schema definitions directly in the schema registry instead of duplicating the schema definitions inline. This is to avoid double maintenance. 
+
+An example directly taken from reference spec is shown below
+
+```yaml
+...
+channels:
+  namespace.source.event.topic-name:
+    ...
+    publish:
+      ...
+      message:
+         $ref: '#/components/messages/topic1Payload'
+components:
+  ...
+  schemas: 
+    ...
+    topic1SchemaValue:
+        schemaFormat: 'application/vnd.apache.avro;version=1.9.0'
+        payload:
+          $ref: https://sit-fdp-pivotal-schema-registry.api.3stripes.io/subjects/pea_fd_fdp.sample.test-value/versions/latest/schema 
+  messages:
+    topic1Payload:
+        $ref: '#/components/schemas/topic1SchemaValue'
+```
+
+**Important note** The used schema is a very simple one, it is only used to illustrate how to refer to it.