Subsequent task adjust parallelism according to CDC source phases

[yeunghl-shoalter]

According to the Flink CDC MySQL connector doc

When the MySQL CDC source is started, it reads snapshot of table parallelly and then reads binlog of table with single parallelism.

To prevent backpressure and maintain ordering of binlog reading, the subsequent task would like to have multiple subtasks in snapshot phase. In binlog reading phase, there shall be only 1 single subtask.

Any idea about adjusting the parallelism of subsequent tasks???