In many use cases such as Spark / Flink ingestion, multiple instances of workers will be instantiated to upload segments in parallel
Currently the default SegmentWriter and SegmentUploader cannot support this easily: One needs to generate slightly different TableConfig and BatchIngestionConfig in order to modify the tmp directory name, segment name and other usages. Otherwise, spawning multiples instances within the same host will cause File write conflict.
Propose to: create a method to easily set parallelism index for SegmentWriter and Uploader; such as adding APIs to modify the parallelism index directly instead of via configuration changes.
CC @npawar, @xiangfu0