This comment needs to be higher up; Amazon has a service for doing just this, dumping 'dumb' files (like json, csv, etc) into S3 buckets and performing SQL queries on them. No need to have to think about how to store things for future querying.

I've used Athena really effectively to solve similar problems. If your data storage is relatively small and/or your queries relatively infrequent, JSON can be a good fit. As one of those dimensions expands, you can decrease costs/increase performance by converting to Parquet and compressing.