Storing and finding AMC results in Amazon S3
In short: AMC can write workflow results as CSV into an S3 bucket in your own AWS account. You create the bucket with the CloudFormation link from instance management. Each workflow gets a folder with sub-folders for daily, weekly or adhoc. AMC also stores the metadata, including the SQL query that was used.
Setup
An AWS account and a bucket are not required for the AMC APIs. They are the way to route results automatically to a place you own. You enter the AWS account and bucket name in the instance, create the bucket in your account with the generated CloudFormation link and allow the AMC account to save results there. According to Amazon, AMC cannot read other files in that bucket.
Without a bucket the only option is the download URL per execution (see the page linked below).
Layout of the path
An example path from the Amazon docs: amc-myinstance-2ofoqzvk/workflow=my_report_daily/schedule=daily/2023-06-03-my_report_daily-ver1.csv.
| Part | Meaning |
|---|---|
| amc-myinstance-2ofoqzvk | Prefix amc-, instance name and a hash that makes the bucket name globally unique |
| workflow=my_report_daily | The workflowId; every saved workflow has its own folder |
| schedule=daily | daily, weekly or adhoc; for schedules aggregationPeriod determines the path |
| 2023-06-03-my_report_daily-ver1.csv | Date of generation, workflow and version number |
Ad hoc executions without a previously saved workflow end up, per the docs, in folders with random names prefixed transient. Next to the CSV file there is a folder with metadata that contains the query that was used.
Processing further
The format is a plain CSV file. Our assessment: any tool that can read CSV from S3 can load these results. In the pages we read, Amazon documents no ready-made connectors to BI tools. How to get the CSV into a spreadsheet or BI tool is described on the page about CSV and BI.