Page 3 of 4.
RDS bills compute, storage, and backups as separate meters, and Multi-AZ silently doubles two of them. Here is what each meter costs and where it leaks.
EBS bills provisioned capacity, provisioned performance and snapshots as three separate meters — and none of them stop when the instance does.
The crossover is 29% average utilization. Below it on-demand wins, above it provisioned does — here is the arithmetic and how to find your own number.
EFS Standard costs 13× S3 per GB-month, so cold data is the whole game — and the lifecycle policy that should move it often silently moves nothing.
CloudWatch bills ingestion, storage, queries, and custom metrics as four separate meters. Here is where the money actually goes, and how to cut each one.
AWS Cost Anomaly Detection is free and catches sudden spend spikes. Here's how to set it up properly — and the kind of waste it will never alert you about.
A working guide to AWS cost optimization: where the money actually leaks, which AWS tools cover what, and a checklist you can run through this week.
A Redshift cluster nobody queries bills the same as one running dashboards. Pausing a business-hours-only warehouse cuts roughly 70% of its compute charge.
An idle load balancer with no healthy targets still bills ~$22/mo. Here's how to find unused ALBs, NLBs, and classic ELBs — and safely delete them.
An idle ElastiCache node bills exactly the same as a busy one — from ~$12/mo per node. How to prove a Redis cluster is unused, then delete it safely.
Unattached EBS volumes and forgotten snapshots bill silently for months. Here's how to find them, delete them safely, and stop the storage from growing back.
Low CPU alone is a bad signal for "idle." Learn how to find EC2 instances nobody uses, what they actually cost, and whether to stop, terminate, or downsize.