Each replication starts where the last one finished, meaning no event gets processed twice.
Performance of the SQL server will in no way be affected by the replication.
There's no limit on how many tables you can replicate - even 200+ tables is okay.
The whole process can be automated using a companion app. If new tables are added to your database, the companion app can fetch these and automatically add them to the replication process. The same applies to the schema changes.