Aurora Storage Over-Provisioning
CostLens flags Aurora DB clusters using only a small fraction of their provisioned storage over 14 days, surfacing ongoing overpayment to review.
Aurora bills for storage, and a cluster persistently using a small fraction of its allocation is paying for capacity it doesn't need. CostLens flags clusters with most of their storage free over the lookback window so you can review the configuration.
How it works
CostLens lists Aurora clusters
Calls rds:DescribeDBClusters and filters to Aurora engines. Aurora Serverless is skipped — it has no fixed allocation.
Reads free storage
Queries the CloudWatch FreeStorageSpace metric (14-day minimum) for each cluster.
Flags over-provisioning
Flags clusters where more than 60% of provisioned storage is free (i.e. under 40% used). Severity scales with the estimated monthly storage waste at ~$0.10/GB.
What to do
- Open the recommendation to see allocated vs. used storage and the estimated monthly waste.
- Review the cluster's storage configuration at the next maintenance window. This check is advisory — Aurora storage cannot be shrunk on a running cluster.
Aurora storage auto-scales
Aurora storage grows automatically and is billed per GB used. This check is a heuristic over reported allocation/free space — use it to spot clusters worth reviewing, not as an exact reclaim figure.
Severity levels
| Severity | Estimated monthly savings |
|---|---|
| critical | >$500 |
| high | $100–$500 |
| medium | $20–$100 |
| low | <$20 |
Required IAM permissions
Read-only — covered by the CostLens ReadOnlyAccess policy from AWS Accounts:
rds:DescribeDBClusters
cloudwatch:GetMetricStatistics
Limitations
- Advisory only — there is no automated fix; address at a storage-config change.
- Aurora Serverless and clusters without a reported allocation are skipped.