Reducing Amazon Aurora Costs by 42% with CloudKeeper

Industry:

Enterprise SaaS

Headquarters:

Menlo Park, California, USA

Founded in:

2020

Company Size:

51 - 200 employees

Featured Tags:

Overview

AccuKnox is a cybersecurity company offering a Zero Trust Cloud-Native Application Protection Platform (CNAPP) to secure cloud environments, applications, Kubernetes workloads, APIs, and AI assets. Its platform supports SaaS, public and private cloud, on-premises, and air-gapped deployments.

AccuKnox’s SaaS architecture uses Amazon Aurora PostgreSQL for database services. Its environment included two Aurora PostgreSQL Serverless v2 clusters, one supporting production workloads and the other demo workloads, with writer and reader instances in each cluster.

Challenges

AccuKnox’s Aurora environment incurred approximately $9,486 in monthly database costs, with opportunities to optimize both compute capacity and storage.

Underutilized reader instances

Reader instances were configured to maintain capacity in line with the writer, despite low CPU utilization and minimal database connections.

High storage I/O costs

Production workloads incurred substantial I/O charges under Aurora Standard storage.

Different workload economics

The demo cluster had lower storage I/O costs, requiring its compute and storage usage to be assessed together before selecting a more cost-effective configuration.

The Solution

Solution

CloudKeeper assessed both Aurora Serverless v2 clusters and implemented two configuration changes to improve cost efficiency without application changes or downtime.

Optimizing Reader Capacity

CloudKeeper changed the reader promotion tiers from Tier 0/1 to Tier 2, allowing reader instances to scale independently of the writer rather than maintaining closely aligned capacity.

This addressed the mismatch between provisioned compute capacity and actual reader utilization while retaining the readers' role as failover targets.

Reducing Storage I/O Costs

CloudKeeper evaluated storage and I/O charges across both clusters and switched their storage configuration from Aurora Standard to Aurora I/O-Optimized.

The change removed separate I/O charges under the I/O-Optimized pricing model. The decision was assessed alongside the reader capacity changes to determine the combined cost benefit for each cluster.

Impact

The configuration changes delivered the following cost reduction, measured from AWS billing by comparing average daily Amazon RDS cost for 22–30 September 2026 with 19–27 October 2026.

Metric	Result
Average daily cost before optimization (22–30 Sep 2026)	$316.20
Average daily cost after optimization (19–27 Oct 2026)	$184.12
Monthly savings ($132.08 per day × 30 days)	$3,962.40
Cost reduction	41.77%
Annual savings (monthly × 12)	$47,548.80
Aurora clusters optimized	2
Downtime during implementation	None

Note: Monthly savings use a 30-day month; annual savings are monthly savings × 12.

Values Delivered Description

Values Delivered

With CloudKeeper, the customer achieved:

Lower database operating costs

Description

Reduced monthly database expenditure by $3,962.40 (41.77%) by optimizing compute and storage configurations.

More efficient compute utilization

Description

Enabled reader instances to scale independently, better aligning capacity with actual demand.

Reduced storage I/O charges

Description

Adopted a pricing model better suited to the combined workload profile of each cluster.

Maintained availability

Description

Implemented the changes without downtime or application modifications while retaining reader instances as failover targets.

Cluster-specific optimization

Description

Evaluated production and demo workloads together across compute and storage costs to identify the appropriate configuration.

Conclusion

CloudKeeper helped AccuKnox identify and address inefficiencies in its Aurora PostgreSQL environment by aligning reader capacity with actual usage and optimizing storage costs. The changes provided a more cost-efficient database configuration while preserving the existing application architecture and failover capability.

With measured savings of 41.77%, about $47,549 a year, AccuKnox gained a practical approach to reducing database expenditure without disrupting its workloads. These figures come from AWS billing for the weeks after the changes.

Talk to our team