== Flink State Processor API | SavepointWriter == I am trying to use State Processor API to cleanup an existing savepoint. The Savepoint size is around 10 TBs. While writing the modified savepoint using SavepointWriter.write(), i am hitting a bottleneck. Wanted to understand the intent behind below piece of code. Why is the parallelism of below sink being forcefully set to 1? flink version : 1.15.1 https://github.com/apache/flink/blob/release-1.15.0-rc1/flink-libraries/flink-stat[…]i/src/main/java/org/apache/flink/state/api/SavepointWriter.java

DataStream<OperatorState> existingOperatorStates = newOperatorStates.getExecutionEnvironment().fromCollection(existingOperators).name("existingOperatorStates");
            existingOperatorStates.flatMap(new StatePathExtractor()).setParallelism(1).addSink(new OutputFormatSinkFunction(new FileCopyFunction(path)));
            finalOperatorStates = newOperatorStates.union(new DataStream[]{existingOperatorStates});