Knowing this could get way out of hand as we go, I'm thinking ahead of time about using some tricks to keep our costs from getting way out of hand, causing us to spend all of Uncle Lonnen's money. I'm thinking about the following: * Setup app, env, project tags in a standard way across all we deploy so we can always see whats running, divvy out permissions, and see the associated costs. * Get automated billing reports funneling into a new S3 bucket for parsing. * Setup a restricted access Netflix Ice instance to chew on said logs and provide us with reporting. * Get in the habit of doing reserved instances for EC2, RDS, and cache servers when we know we'll have a piece of infra for at least a year.
We're waiting on Corey Shields and crew to determine how they want ot provision read only IAM ids/creds.
The AWS accounts already have their cost management in place, this is no longer desired.