...

- The change made in this proposal is both source backward-compatible and binary backward compatible. Their code can compile and run correctly without change.
- For users who want to enable partition expansion for its input streams, they can do the following:
  - Set grouper class to GroupByPartitionWithFixedTaskNum if the job is using GroupByPartition as grouper class
  - Set grouper class to GroupBySystemStreamPartitionWithFixedTaskNum if the job is using GroupBySystemStreamPartition as grouper class 
  - Change their custom grouper class implementation to extend the new interface if the job is using a custom grouper class implementation.
  - Set job.coordinator.monitor-partition-change to true in the job configuration
  - Run ConfigManager

Rejected Alternatives

1. Allow task number to increase instead of creating a new grouper class.

Allow task number to increase is useful since it increases the performance of a given Samza job. However, this feature alone does not solve the problem of allowing partition expansion. For example, say we have a job that joins two streams both of which have 3 partitions. If partition number of one stream increases from 3 to 6, we would still want the task number to remain 3 to make sure that messages with the same key from both streams will be handled by the same task. This needs to be done with the new grouper classes proposed in this doc.

2. Add the new method to the existing interface SystemStreamPartitionGrouper instead of creating a new interface.

The advantage of this alternative solution is that it requires less interface class and shorter class hierarchy. But we choose to follow the existing pattern of interface extension in Samza as we do with BalancingTaskNameGrouper.

3. Add new config and class for user to specify the new-partition to old-partition mapping strategy based on the input system.

The current proposal relies on the input system to meet the specified operational requirements. While these requirements can be satisfied by Kafka, they may or may not be satisfied by other systems such as Kinesis. We can support partition expansion for more input systems than Kafka if user is able to express the new-partition to old-partition mapping strategy. However, since we currently don't know how user is going to use such config/class to do it, we choose to keep the current SEP simple and only add new config/class when we have specific use-case for them.

4. 2. Use additional repartitioning stage to repartition data from input stream to another internal stream of the old partition count

...

Future work

1. Enable task number expansion in Samza. Allow task number to increase instead of creating a new grouper class.

Allow task number to increase is useful since it increases the performance of a given Samza job. However, this feature alone does not solve the problem of allowing partition expansion. For example, say we have a job that joins two streams both of which have 3 partitions. If partition number of one stream increases from 3 to 6, we would still want the task number to remain 3 to make sure that messages with the same key from both streams will be handled by the same task. This needs to be done with the new grouper classes proposed in this doc.

2. Add 2. Add new config and class for user to specify the new-partition to old-partition mapping strategy based on the input system. 

The current proposal relies on the input system to meet the specified operational requirements. While these requirements can be satisfied by Kafka, they may or may not be satisfied by other systems such as Kinesis. We can support partition expansion for more input systems than Kafka if user is able to express the new-partition to old-partition mapping strategy. However, since we currently don't know how user is going to use such config/class to do it, we choose to keep the current SEP simple and only add new config/class when we have specific use-case for them.