m90
Right now, the script is writing the database dump into a temp file: https://github.com/m90/mongo-lambda-backup/blob/0f7d3d0c78f967e560c18b075679cc7559db65e5/mongo_lambda_backup/handler.py#L55-L58 In certain scenarios (when hitting certain AWS Lambda limits) it would be a nice addition if the content could instead be buffered in memory before being written to S3.