Apply suggestions from code review

Co-authored-by: Oliver Howell <[email protected]>

docs/modules/integrate/pages/cdc-connectors.adoc

@@ -58,7 +58,7 @@ PostgreSQL-specific connector:
 </dependency>
 ----
 
-== CDC as a Source
+== CDC as a source
 
 The Java API supports the following types of CDC source:
 
@@ -244,7 +244,7 @@ Each database type has its own database type-to-struct type mappings. For specif
 |STRING|-|String
 
 The `RecordPart#value` field contains Debezium's message in a JSON format. This JSON format uses string as date representation,
-instead of ints, which are standard in Debezium, but harder to deal with.
+instead of ints, which are standard in Debezium but harder to work with.
 
 [NOTE]
 ====

docs/modules/integrate/pages/legacy-cdc-connectors.adoc

@@ -68,7 +68,7 @@ depending on the DB configuration, can be relatively short, so if the CDC
 source has to replay data for a long period of inactivity, then there
 can be data loss. With careful management, the _at least once_ guarantee can be practically implemented.
 
-== CDC as a Sink
+== CDC as a sink
 
 Change data capture is a source-side functionality in Jet, but Hazelcast also
 offers specialized sinks that simplify applying CDC events to a map, which gives you the ability to reconstruct the contents of the original database table. The sinks expect to receive `ChangeRecord`

docs/modules/pipelines/pages/cdc-database-setup.adoc

@@ -255,7 +255,7 @@ In our tests we didn't manage to make it output much more than 20,000
 records/second, so on a powerful server running the database it
 shouldn't affect normal operations too severely.
 
-==== Failure Tolerance
+==== Failure tolerance
 
 PostgreSQL's failure tolerance associated with replication slots is
 somewhat lacking in certain aspects. The CDC connector can quite nicely