max parallelism is static to 128 unless you pass in write.parallelism

[ryanvanhuuksloot]

I was expecting with #203 that if your dataset was in region us then you'd be able to use up to the new cap of 512 without specifically setting the write.parallelism

What I found was that you need to pass write.parallelism
Even adding write.region wasn't enough to use a higher parallelism than 128 (without the write.parallelism)

[jayehwhyehentee]

128 is the Flink default. The connector enforces a max parallelism, but does not set the current job's parallelism. Passing write.parallelism is the expected behavior here.
Btw, I gope you're setting the region here (for datastream API) or here (for table API).