Definitely not about the actual data size 😆 This is regarding the actual schema definition, in Airbyte it is stored as a JSON that defines the expected columns, types and other structure of the data. If you have many columns (usually 200+) you may get an I/O error when trying to set up the connection:
https://github.com/airbytehq/airbyte/issues/12257
This issue is
currently being addressed and hopefully performance will be improved soon. It basically comes down to a hardcoded 4MB message limit configured for gRPC in temporal, a dependency for airbyte