Reducing Monthly Cloud Costs by 28% Through Compute and Storage Optimization
The client’s cloud environment had grown steadily as new workloads, storage, and backup data were added. Although applications remained stable and there were no major operational issues, monthly infrastructure costs continued to rise. A review showed that some long-running workloads were still using on-demand pricing, while interruptible jobs ran on full-priced compute and older data remained in standard storage.
Bobcares reviewed compute usage, workload behavior, and storage lifecycles to identify lower-cost options. Predictable workloads were moved to Reserved Instances, suitable background jobs to Spot-based autoscaling groups, and older data to lower-cost archive tiers.
Business Challenge
The client's cloud bill was increasing despite stable application performance and workloads. Long-running instances used on-demand pricing, flexible background jobs ran on full-priced compute, and older backups and logs remained in standard storage. The client needed to reduce spending without disrupting production workloads or affecting end users.
Technology Used
Blueprint
Compute usage was first separated by workload behavior. Predictable application and database workloads were matched with Reserved Instances, while interruptible background jobs were placed in Spot-based autoscaling groups. A storage audit then identified older backups and logs for movement into lower-cost archive tiers through lifecycle policies.
Key Outcomes
- Monthly cloud spending was reduced by approximately 28%.
- Predictable workloads received more suitable pricing through Reserved Instances.
- Flexible background processing moved to lower-cost Spot capacity.
- Older backups and logs were moved to lower-cost archive tiers.
- Production stability was maintained throughout the optimization.
- No downtime or major migration work was required.
- End users experienced no impact from the changes.
Customer Profile
The client operated a growing cloud environment supporting production workloads, background processing, backups, and stored logs. Infrastructure usage had expanded over time as additional workloads and data were added. The applications remained stable, but the monthly cloud bill continued to increase. A review was needed to determine whether existing resources were still using appropriate pricing and storage options.
The Challenges
- Long-running compute instances continued to use on-demand pricing despite stable usage patterns.
- Background processing jobs were running on full-priced instances even though they could tolerate interruptions.
- Backup data and archived logs remained in standard storage despite limited access.
- Storage requirements continued to grow as backup retention increased.
- Cloud spending increased without a corresponding production performance problem.
- Cost reductions had to be achieved without disrupting application availability or user experience.
Why Bobcares?
- Reviewed actual workload behavior rather than simply reducing resource capacity.
- Distinguished predictable workloads from processes that could tolerate interruptions.
- Used existing retry behavior to identify suitable candidates for Spot capacity.
- Examined storage access patterns before changing storage tiers.
- Focused on pricing and lifecycle changes that did not require major architectural modifications.
- Preserved production stability while addressing the sources of unnecessary spending.
What We Delivered
- Reviewed compute usage and identified continuously running instances with stable workload patterns.
- Shifted suitable long-running workloads to Reserved Instances.
- Identified background processing workloads that could tolerate compute interruptions.
- Moved those workloads to Spot-based autoscaling groups.
- Reviewed backup and log storage based on age and access patterns.
- Introduced lifecycle policies for older data.
- Moved historical backups and logs into lower-cost archive storage tiers.
- Monitored the environment throughout the optimization to ensure production workloads remained stable.
Results
Monthly Cloud Cost
- Before: Monthly spending continued to increase as compute, storage, and backup requirements grew.
- After: The combined optimization reduced the overall monthly cloud bill by approximately 28%.
Compute Pricing
- Before: Stable, continuously running workloads used on-demand pricing.
- After: Suitable long-running workloads used Reserved Instances for more efficient pricing.
Background Processing
- Before: Interruptible background jobs ran on full-priced instances.
- After: Suitable jobs ran through Spot-based autoscaling groups and continued to handle interruptions through their existing retry behavior.
Storage Costs
- Before: Older backups and archived logs remained in standard storage despite limited access.
- After: Lifecycle policies moved older data into lower-cost archive tiers.
Production Stability
- Before: Cost reduction had to be achieved without introducing operational risk.
- After: The optimization was completed without downtime, major migration work, or an impact on end users.
The Business Impact
- Reduced recurring cloud expenditure without changing core application behavior.
- Improved the cost efficiency of stable production workloads.
- Lowered the expense of running interruption-tolerant background processing.
- Reduced the long-term cost of retaining historical backups and logs.
- Avoided downtime and major infrastructure migration during the optimization.
- Created a clearer relationship between workload characteristics and the pricing model used.
Ongoing Collaboration
- Review compute usage as workload patterns change.
- Reassess Reserved Instance suitability for stable workloads.
- Monitor Spot-based workloads and their interruption behavior.
- Review storage age and access patterns as data accumulates.
- Maintain lifecycle policies for older backups and logs.
- Monitor monthly cloud spending for new optimization opportunities.
Conclusion
The client’s rising cloud bill was not caused by a single infrastructure problem. It came from pricing models and storage choices that no longer matched how workloads were being used. Bobcares aligned stable workloads with Reserved Instances, flexible processing with Spot capacity, and older data with archive storage, reducing monthly cloud spending by approximately 28% without disrupting production.
