In order to collect statistics for downloads/uploads we send the events to kinesis firehose streams. The stream will funnel the data to S3 that is later analyzed with AWS Athena.
For this task we need to refactor the stack builder so that the streams are created with potentially support for Apache Parquet so we can query them easily with AWS Athena. We also need to integrate with Glue so that the table schema is setup for the stream.
Some refactoring is needed in the synapse backend so that sending the download/uploads (and in general other type of records) to kinesis firehose streams is robust enough.