* Add in-process caching to dynamodb collector
* Add metrics for dynamodb consumed capacity and report size
* Log and return errors during report collection
* Increase compression to the max
* Put reports in S3 and just use DynamoDB as an index.
* Review feedback