How do I reduce Aurora PostgreSQL costs on AWS?

The biggest way to reduce RDS costs for Aurora PostgreSQL is consolidating cluster sprawl — a common FinOps anti-pattern where each microservice runs its own $208/mo Aurora cluster. CostPatrol, an AWS cost optimization tool, found 17 such clusters costing $3,549/mo that could be consolidated into 2 shared clusters, saving $2,192-$3,132/mo. Consolidation involves using pg_dump to export each database, creating separate schemas in a shared cluster, and deleting the originals. This is the single biggest Aurora PostgreSQL pricing optimization most teams miss.

What is RDS cluster sprawl and how much does it cost?

RDS cluster sprawl is a FinOps anti-pattern where each microservice runs its own dedicated database cluster instead of sharing a cluster with separate schemas. Each Aurora PostgreSQL cluster has a minimum pricing floor of ~$208/mo regardless of usage. In this real AWS cost optimization scan, 17 clusters across 4 regions — many with zero active connections — cost $6,496/mo when 2 shared clusters would handle the same workloads. Detecting and consolidating RDS sprawl is one of the highest-impact ways to reduce your AWS bill.

What is the best way to find and fix RDS cluster sprawl?

The best way to find RDS cluster sprawl is using an AWS cost optimization tool like CostPatrol that automatically detects multiple Aurora clusters in the same region running the same engine. CostPatrol identifies clusters that could be consolidated, checks for idle instances with zero connections, and provides exact cluster names, per-cluster Aurora PostgreSQL pricing, and step-by-step consolidation commands (pg_dump export, schema creation, cluster deletion). It's more effective than AWS Trusted Advisor, which doesn't detect sprawl patterns or provide consolidation plans.

Real scan — anonymized data

CostPatrol found $6,496/mo
in savings — across 4 AWS regions

We scanned a production SaaS account across us-east-1, us-east-2, eu-west-2, and us-west-2. 56 optimization opportunities across 4 rules — dominated by database cluster sprawl.

By Mohamed Shehabeldin, Founder of CostPatrol

Total savings

$6,496/mo

$3,879 - $6,496

Findings

Rules matched

Regions scanned

Every finding, with exact fix commands

56 optimization opportunities across 4 rules and 4 regions. Real dollar amounts. Anonymized resource names.

RDS Cluster Sprawl

RDS-O005 · 17 clusters across 2 regions

$2,192 - $3,132/mo

9x Aurora PostgreSQL clusters (us-east-1)

$1,169 - $1,670/mo

9 clusters · $1,879.20/mo total · Each microservice running its own dedicated cluster

→ Consolidate 9 Aurora PostgreSQL clusters into 1 shared cluster with separate schemas

svc_auth_db_production
svc_billing_db_production
svc_catalog_db_production
svc_notify_db_production
svc_orders_db_production
svc_search_db_production
svc_users_db_production
svc_inventory_db_production
svc_payments_db_production

# 1. Create final snapshots of each cluster $ aws rds create-db-cluster-snapshot \ --db-cluster-identifier svc-auth-dbcluster-prod \ --db-cluster-snapshot-identifier svc-auth-final-snapshot # 2. Export data using pg_dump $ pg_dump -h svc-auth-dbcluster-prod.cluster-xxxx.us-east-1.rds.amazonaws.com \ -U admin -d svc_auth_db_production > svc_auth_backup.sql # 3. Import into shared cluster as separate schema $ psql -h shared-cluster-prod.cluster-xxxx.us-east-1.rds.amazonaws.com \ -U admin -c "CREATE SCHEMA svc_auth;" $ psql -h shared-cluster-prod.cluster-xxxx.us-east-1.rds.amazonaws.com \ -U admin -d shared_production < svc_auth_backup.sql # 4. Repeat for all 9 clusters, then delete originals

8x Aurora PostgreSQL clusters (us-east-2)

$1,023 - $1,462/mo

8 clusters · $1,670.40/mo total · Same pattern: one cluster per microservice

→ Consolidate 8 Aurora PostgreSQL clusters into 1 shared cluster with separate schemas

svc_analytics_db_production
svc_shipping_db_production
svc_pricing_db_production
svc_reviews_db_production
svc_checkout_db_production
svc_messaging_db_production
svc_reports_db_production
svc_support_db_production

# Same consolidation process as above for us-east-2 clusters $ aws rds create-db-cluster-snapshot \ --db-cluster-identifier svc-analytics-dbcluster-prod \ --db-cluster-snapshot-identifier svc-analytics-final-snapshot # Repeat snapshot → export → import → delete for all 8 clusters

Idle RDS Instances

RDS-O001 · 16 resources across 4 regions

$1,670 - $3,341/mo

svc-auth-dbcluster-prod-1

$104 - $209/mo

us-east-1 · Aurora PostgreSQL · $208.80/mo · Zero active connections

→ Stop, downsize, or delete low-utilization RDS instance

# Option 1: Snapshot and delete $ aws rds create-db-cluster-snapshot \ --db-cluster-identifier svc-auth-dbcluster-prod \ --db-cluster-snapshot-identifier svc-auth-dbcluster-prod-final $ aws rds delete-db-instance --db-instance-identifier svc-auth-dbcluster-prod-1 --skip-final-snapshot # Option 2: Stop instance $ aws rds stop-db-instance --db-instance-identifier svc-auth-dbcluster-prod-1

svc-analytics-dbcluster-prod-1

$104 - $209/mo

us-east-2 · Aurora PostgreSQL · $208.80/mo · Zero active connections

→ Stop, downsize, or delete low-utilization RDS instance

$ aws rds stop-db-instance --db-instance-identifier svc-analytics-dbcluster-prod-1

svc-fulfillment-dbcluster-prod-1

$104 - $209/mo

eu-west-2 · Aurora PostgreSQL · $208.80/mo · Zero active connections

→ Stop, downsize, or delete low-utilization RDS instance

$ aws rds stop-db-instance --db-instance-identifier svc-fulfillment-dbcluster-prod-1

svc-media-dbcluster-prod-1

$104 - $209/mo

us-west-2 · Aurora PostgreSQL · $208.80/mo · Zero active connections

→ Stop, downsize, or delete low-utilization RDS instance

$ aws rds stop-db-instance --db-instance-identifier svc-media-dbcluster-prod-1

and 12 more idle RDS instances across 4 regions — $1,253 - $2,506/mo in additional savings

Previous-Generation Instance Types

EC2-O001 · 20 resources across 4 regions

$14 - $21/mo

i-0a1b2c3d4e5f0001

$0.69 - $1.04/mo

us-east-1 · t2.micro · stopped · $8.35/mo

→ Upgrade to t3.micro (better price/performance, Nitro-based)

$ aws ec2 stop-instances --instance-ids i-0a1b2c3d4e5f0001 $ aws ec2 modify-instance-attribute --instance-id i-0a1b2c3d4e5f0001 --instance-type '{"Value":"t3.micro"}' $ aws ec2 start-instances --instance-ids i-0a1b2c3d4e5f0001

i-0a1b2c3d4e5f0010

$0.69 - $1.04/mo

us-east-2 · t2.micro · stopped · $8.35/mo

→ Upgrade to t3.micro (better price/performance, Nitro-based)

$ aws ec2 modify-instance-attribute --instance-id i-0a1b2c3d4e5f0010 --instance-type '{"Value":"t3.micro"}'

i-0a1b2c3d4e5f0018

$0.69 - $1.04/mo

eu-west-2 · t2.micro · stopped · $8.35/mo

→ Upgrade to t3.micro (better price/performance, Nitro-based)

$ aws ec2 modify-instance-attribute --instance-id i-0a1b2c3d4e5f0018 --instance-type '{"Value":"t3.micro"}'

and 17 more previous-generation t2.micro instances across 4 regions

Lambda Memory Right-Sizing

LAM-O002 · 18 resources across 4 regions

$2 - $3/mo

svc-inventory-handler-prod

$0.08 - $0.11/mo

us-east-1 · 1024MB allocated · 190MB peak usage · $0.14/mo

→ Reduce memory from 1024MB to 256MB (75% reduction, still 35% headroom)

$ aws lambda update-function-configuration \ --function-name svc-inventory-handler-prod \ --memory-size 256

svc-reports-handler-prod

$0.36 - $0.45/mo

us-east-2 · 1024MB allocated · $0.90/mo

→ Reduce memory from 1024MB to 256MB

$ aws lambda update-function-configuration \ --function-name svc-reports-handler-prod \ --memory-size 256

svc-support-handler-prod

$0.34 - $0.43/mo

us-east-2 · 1024MB allocated · $0.86/mo

→ Reduce memory from 1024MB to 256MB

$ aws lambda update-function-configuration \ --function-name svc-support-handler-prod \ --memory-size 256

and 15 more over-provisioned Lambda functions across 4 regions

What AWS Cost Explorer misses

AWS Cost Explorer

✕ Shows $3,549/mo in RDS spend — with no per-cluster breakdown

✕ Tells you "RDS costs went up" — not which clusters are idle

✕ Cannot detect that 17 clusters could be consolidated into 2

✕ No consolidation plan — you're on your own

CostPatrol

✓ Exact cluster name, region, utilization, and savings per instance

✓ "17 Aurora clusters — each microservice has its own $208/mo cluster"

✓ Consolidation plan: merge 9 into 1, merge 8 into 1

✓ Step-by-step commands to snapshot, export, and consolidate

AWS tells you RDS costs went up. CostPatrol tells you that 17 microservice databases — each running their own $208/mo Aurora cluster — could be consolidated into 2, saving up to $3,132/mo.

More real AWS scan reports

$1,112/mo

Full-stack waste across 12 rules

35 findings · 12 rules · 7 regions

$284/mo

Hidden single-region waste

14 findings · 6 rules · 1 region

See what CostPatrol finds in your AWS account

Free scan shows your total savings. Upgrade to Pro for full findings, fix commands, and daily Slack alerts.

Scan my account free

CostPatrol found $6,496/mo in savings — across 4 AWS regions

Every finding, with exact fix commands

What AWS Cost Explorer misses

More real AWS scan reports

See what CostPatrol finds in your AWS account

CostPatrol found $6,496/mo
in savings — across 4 AWS regions