Right now the tidb2dw will connect to upstream TiDB to fetch some information like schema and column default value which break the simple dependency principle. The tidb2dw service should only have dependcy on the shared object storage like S3 and GCS, in this way, there should not have upstream database permission issue and network dependency issues since upstream clusters might in different VPC networks.
Right now the tidb2dw will connect to upstream TiDB to fetch some information like schema and column default value which break the simple dependency principle. The tidb2dw service should only have dependcy on the shared object storage like S3 and GCS, in this way, there should not have upstream database permission issue and network dependency issues since upstream clusters might in different VPC networks.