Custom partition key for the record that aggregates user records

[ruslandanilin]

Hello,

Currently, when user records with random partition keys are aggregated - they are sent to different shards as expected.
But the partition key of a "parent" / "container" record always equals "a". Thus, it's not possible to effectively use AWS Lambda Kinesis event source with a parallelization option called "Concurrent batches per shard" > 1, because records with the same partition key are read sequentially.

Is it possible to set a random partition key for the "parent" / "container" record?

Thank you