CloudThinker Use Case: Global AWS FinOps & Cost Optimization
Nobody remembered who launched the m5.xlarge in ap-southeast-1.
It had been running for nine months. CPU utilization: 0.2 percent. Memory usage: negligible. Network traffic: the occasional health check ping. It was a ghost, a server doing nothing, costing $120 every month, part of a $4,350 monthly AWS bill that no one could fully explain.
Linh, the operations lead at a mid-size SaaS company we will call PeakStack, discovered the instance on a Tuesday afternoon. She was not looking for it. She was trying to answer a simpler question: why had their AWS bill increased 20 percent over the past quarter despite no new product launches?
The answer, it turned out, was not one thing. It was thirty-one things.
The Audit That Changed Everything
PeakStack's AWS environment had grown the way most do: organically. A developer spins up an instance for a proof of concept. A database gets provisioned "just in case." Elastic IPs get allocated during a migration and never released. Nobody intends to waste money. But across 21 AWS regions, waste accumulates like sediment.
Linh had tried manual audits before. She would spend a Friday afternoon clicking through the EC2 console, region by region, checking instance utilization. By the time she reached ap-south-1, she was exhausted and it was 6 PM. The next Friday, she would start over because the environment had changed.
When she connected CloudThinker to PeakStack's AWS accounts, the platform completed what she could not: a comprehensive scan across all 21 regions, every service, every resource.
The results were sobering.
Discovery Results
- 107 AWS resources inventoried and analyzed
- 31 optimization opportunities identified
- Potential savings: $1,956/month (~$23,500/year)
Thirty-one opportunities. Not theoretical scenarios from a best-practices checklist, but specific resources in specific regions with specific dollar amounts attached.
The Ghost and Its Friends
The m5.xlarge in ap-southeast-1 was just the most dramatic find. Linh started calling it "the ghost" because it had no owner, no purpose, and no one noticed when CloudThinker eventually recommended shutting it down.
But the ghost had company:
Storage waste was hiding in plain sight. Two EBS volumes totaling 508 GB sat unattached to any instance. They had been orphaned during a migration eight months earlier. Cost: $50.80 per month for storage no one was using. S3 buckets across multiple accounts had no Intelligent-Tiering or lifecycle policies, meaning data accessed once a year sat in the same storage class as data accessed hourly. Enabling automated tiering would save an estimated $300 to $400 per month.
Compute was chronically oversized. Beyond the ghost instance, CloudThinker's analysis revealed a pattern: production workloads running on instance types chosen during initial setup and never revisited. Reserved Instances and Spot Instances could cover stable and fault-tolerant workloads respectively, saving an additional $400 and $150 per month.
Network costs were leaking quietly. Six Elastic IPs sat unassociated with any instance, each incurring charges. Small individually at $3.60 per month each, but symptomatic of a larger problem: resources provisioned and forgotten.
Quantified Results
| Metric | Current | Optimized | Savings |
|---|---|---|---|
| Monthly Cost | $4,350 | $3,261 | $1,089 (25%) |
| Annual Cost | $52,200 | $39,132 | $13,068 |
| ROI Timeline | - | - | Immediate–3 months |
A 25 percent reduction. Not from a massive re-architecture or a cloud migration project, but from finding and fixing what was already there.
The Implementation Playbook
Linh did not try to fix everything at once. She organized the 31 opportunities into three phases, starting with the changes that required the least effort and carried the least risk.
Quick Wins (Week 1-2)
- Release 6 unassociated Elastic IPs: $21.6/month saved
- Delete 2 idle EBS volumes: $50.8/month saved
- Enable S3 Intelligent-Tiering: $300-400/month saved
These changes took less than a day and required no downtime. The Elastic IPs were released in minutes. The EBS volumes, after confirming they were truly orphaned, were deleted the same afternoon. S3 Intelligent-Tiering was enabled with a single policy change per bucket.
Medium-term (Month 1-3)
- Rightsize EC2 instances: $120/month saved
- Implement Reserved Instances for stable workloads: $400/month saved
- Clean up old snapshots: $50/month saved
These required more analysis. Rightsizing meant monitoring the replacement instances for a week to confirm performance was unaffected. Reserved Instance purchases required forecasting workload stability over the commitment period.
Long-term (Month 3-6)
- Adopt Savings Plans strategically: $200/month saved
- Implement Spot Instances for batch workloads: $150/month saved
- Optimize cross-region data transfer: $100/month saved
The longer-term optimizations involved architectural decisions: which workloads could tolerate Spot Instance interruptions, which regions should host which data, and how to structure Savings Plans for maximum coverage without overcommitment.
Top 3 Highest ROI Actions
1. S3 Intelligent-Tiering
- Effort: Low (1-2 hours)
- Savings: $300-400/month
- Risk: None (AWS-managed)
- ROI: Immediate
2. EC2 Rightsizing
- Effort: Medium (4-8 hours)
- Savings: $120/month
- Risk: Low (requires monitoring)
- ROI: 1 week
3. Reserved Instances
- Effort: Medium (planning required)
- Savings: $400/month
- Risk: Low (commitment-based)
- ROI: Immediate after purchase
The All-Hands That Told a Different Story
Six weeks after the initial scan, Linh presented at PeakStack's monthly all-hands. She showed two numbers: $4,350 (what they had been spending) and $3,261 (what they were spending now). But the story she told was not about the savings. It was about visibility.
"We had 107 resources across 21 regions," she told the team. "Before CloudThinker, I could audit maybe two regions in an afternoon. Now I can see all of them, all the time. The ghost instance in Singapore was running for nine months. That does not happen anymore."
The engineering team's reaction was telling. Nobody was embarrassed about the waste. They were relieved that someone had finally built a system to catch it. The provisioning habits did not change overnight, but the visibility did. Developers started checking CloudThinker's recommendations before spinning up new resources. The monthly review meeting went from a blame session to a planning session.
Annual savings: $13,068. Time saved: 40+ hours per month in manual analysis. But the real outcome was simpler than any metric: PeakStack finally understood where their money was going.
CloudThinker transforms cloud cost management from reactive firefighting to proactive financial engineering.