12 posts tagged with "aws"

Short version: With provisioned DynamoDB capacity you pay for the read/write units you reserve, used or not. Tables are routinely over-provisioned "to be safe" and then run at 5-10% utilization for years. Here is how to find them and pick the right fix: right-size, turn on auto-scaling, or move to on-demand.

Short version: By default, a CloudWatch log group keeps logs forever. Stored logs cost $0.03/GB-month, and every Lambda, ECS task, and API Gateway you have been running for years has been piling them up. Setting a retention policy is a one-line fix per log group. Here is how to find the never-expiring ones and cap them.