Results

Situation: already partly optimized in-house, but with plenty still on the table across a large, EDP-covered footprint.

What we did: decommissioned unused RDS databases found in the first week (~$378k/yr), converted mis-provisioned EBS IO1 volumes to GP3 (~$156k/yr), moved steady-usage Aurora Serverless to provisioned (~$125k/yr), rightsized RDS and covered the remaining footprint with Reserved Instances (~$161k/yr), rightsized ElastiCache and migrated it to Valkey (~$86k/yr), removed unattached EBS volumes and old snapshots (~$27k/yr), plus cross-region RI cleanup and CloudWatch log-ingestion tuning.

Fast payback: the first two weeks alone delivered ~$378k/year in savings, enough to cover the entire cost of the engagement within a few months.

Result: more than $1.1M/year in savings on the bill, and counting.

Situation: a bill growing with the business, from $30k to $50k/month over a year, already optimized in places.

What we did: RDS (~40% of the bill) stopped over-using Aurora I/O-Optimized, rightsized, moved to Serverless v2 where it fit and bought RIs (~30% off); EC2 and Fargate (~30%) got Savings Plans covering ~70% of capacity plus task and instance rightsizing (~40% off); ElastiCache rightsizing, RIs and Valkey (~60% off); S3 storage-tier selection (~63% off); EBS IO1 to GP3 (~90% off in-scope volumes); removed avoidable cross-AZ traffic.

Result: cut to $36k/month, about $14k/month or ~$168k/year, roughly 30% of the entire bill.