Add "first-class" support to pull and publish table and column comments. [DBZ-2237]

Migrated from [DBZ-2237](https://issues.redhat.com/browse/DBZ-2237)

We are copying the data from our MySQL database into Snowflake and would like to also copy the comments for each table's columns.

This information is currently published to the _schema-history_ topic as raw DDL/ALTER statements.  However, the information is difficult to process given its unstructured / textual nature (see below).

Could a better mechanism be added to the framework to add DDL comments as a "first-class" field?  That is, provide a pre-determined place for the data and make it consistent across database engines.  For me, the MVP would be to add this for MySQL and then subsequently for the other supported databases.

 
{code:java}
"ddl" : "Alter table `dvm`.`brand` \r\n change `id` `id` int(11) NOT NULL Auto_increment COMMENT 'the id for the record'" 
{code}
 

Start by reviewing how MySQL DDL and schema-history events are published, especially the raw ALTER statements containing table and column comments. Define the proposed first-class comment field and its consistency requirements, beginning with the MySQL MVP. Done means comments are available in a structured, predictable form rather than only as unstructured DDL text.

Written by the indexing model from the issue text.