Add documentation for Partitions initial offsets (#996)

* Add documentation for Partitions initial offsets

---------

Co-authored-by: rebekah-lawrence <[email protected]>

docs/modules/integrate/pages/kafka-connector.adoc

@@ -75,6 +75,12 @@ committing using `enable.auto.commit` and configuring
 link:https://kafka.apache.org/22/documentation.html[Kafka documentation]
 for the descriptions of these properties.
 
+You can also explicitly specify exact initial offsets for the Kafka source using https://docs.hazelcast.org/docs/{full-version}/javadoc/com/hazelcast/jet/kafka/KafkaSources.html#kafka(java.util.Properties,com.hazelcast.function.FunctionEx,com.hazelcast.jet.kafka.TopicsConfig)[`TopicsConfig` parameter^].
+Note that initial offsets provided in `topicConfig` will always have priority over offsets stored in Kafka or associated with a given consumer group.
+Those offsets are used only when the job is started for the first time after submission.
+Afterwards, the regular fault tolerance mechanism described above is used. 
+This option is not supported when processing guarantees are disabled.
+
 == Transactional Guarantees
 
 As a sink, the Kafka connector provides exactly-once guarantees at the cost of using
@@ -149,4 +155,4 @@ This connector supports heterogeneous messages. For example, say you have these
 
 If you map the column `petName`, it will have the value `null` for the
 entry with `key=1`. This scenario is supported. Similar behavior works
-with Avro format.
+with Avro format.