schema_only mode seems to be slow when replicating database with a lot of tables.

[subkanthi]

There is a debezium configuration parameter schema.history.internal.store.only.captured.tables.ddl which should only include the ddl for the tables listed in the parameter.
But during testing it looks like a lot of time is spent during startup to retrieve the schema for all the tables.

Implementation:
Try to create a database with 15k tables with the configuration variables set in debezium.

[subkanthi]

The following configuration works in limiting debezium to only snapshot the tables listed in the table.include.list

table.include.list: "sbtest.table1,sbtest.table2,sbtest.table3"
schema.history.internal.store.only.captured.tables.ddl: "true"
schema.history.internal.store.only.captured.databases.ddl: "true"

Regex:

"table.include.list","employees.table[1-3]$"

[debezium-mysqlconnector-embeddedconnector-change-event-source-coordinator] INFO io.debezium.relational.RelationalSnapshotChangeEventSource - Snapshot step 3 - Locking captured tables [employees.table1, employees.table2, employees.table3]

Before configuration with 15k tables:

[output_before.log](https://github.com/Altinity/clickhouse-sink-connector/files/14749835/output_before.log)

After configuration: limiting to 3 tables:
output.log

[debezium-mysqlconnector-embeddedconnector-change-event-source-coordinator] INFO io.debezium.relational.RelationalSnapshotChangeEventSource - Snapshot step 3 - Locking captured tables [sbtest.table1, sbtest.table2, sbtest.table3]