Configure PostgreSQL EC2 machine to eliminate bandwidth cost for downloading DB backups

[MikeTheCanuck]

Problem

Currently we take the .sql.gz backup files from the S3 hacko-data-archive bucket, download them to the root volume of the PostgreSQL EC2 VM and restore them to stand up a new database. [This is the solution enabled by #17]

This comes with a cost - AFAIK, AWS charges us for the bandwidth consumed by the download of each file out of S3 to another device "on the Internet".

Since we're not yet routing S3 > EC2 traffic through AWS' backbone, this is costing us non-zero dollars (not much, but let's not waste $$).

Solution

• configure an S3 Endpoint in the EC2 UI Create an S3 Endpoint in the current VPC #39
• add the routing changes needed to make the traffic route over the link provided by the Endpoint
• add an appropriate IAM role to grant the EC2 VM access via awscli to the S3 bucket Role needed to assign to our PostgreSQL EC2 machine for S3 access #41