AWS Backup Cost Analysis
EBS snapshot costs were growing month-over-month with no clear explanation or optimization strategy.
The Problem
An enterprise workload’s AWS Backup costs were growing steadily month-over-month. The team knew snapshots were accumulating but didn’t have visibility into:
- Why costs were increasing at that rate
- Whether the backup configuration matched actual recovery needs
- What steady-state costs would look like
- How to optimize without compromising recovery capabilities
The Solution
Performed a comprehensive cost analysis that revealed:
- 57 recovery points per volume from overlapping daily/weekly/monthly/yearly backups
- No cold storage tiering — all backups stored at warm tier pricing ($0.05/GB vs $0.0125/GB)
- ~$38,000/year potential savings by enabling cold storage transitions
How It Works
Current Cost Breakdown
| Workload | Monthly Cost | Annual Cost |
|---|---|---|
| Production | ~$2,300 | ~$27,600 |
| Secondary | ~$2,660 | ~$31,900 |
| Total | ~$4,960 | ~$59,500 |
Backup Plan Configuration
The backup plans create 57 recovery points per volume:
| Backup Type | Frequency | Retention | Recovery Points |
|---|---|---|---|
| Daily | 1/day | 30 days | 30 |
| Weekly | 1/week | 90 days | ~13 |
| Monthly | 1/month | 365 days | 12 |
| Yearly | 1/year | 730 days | 2 |
| Total | ~57 |
Steady State Formula
Developed a formula to calculate expected storage at steady state:
Steady State Storage = Volume Size Ă— Multiplier
Where Multiplier = 1 + (Change Rate Ă— (Recovery Points - 1))
Example:
- Volume Size: 500 GB
- Change Rate: 5%
- Recovery Points: 57
- Multiplier: 1 + (0.05 Ă— 56) = 3.8
- Steady State: 500 GB Ă— 3.8 = 1,900 GB per volume
The Fix: Cold Storage Tiering
Current configuration (no cold storage):
{
"DeleteAfterDays": 365
}
Recommended configuration:
{
"MoveToColdStorageAfterDays": 14,
"DeleteAfterDays": 365
}
Cost Impact
| Scenario | Monthly Cost | Annual Cost |
|---|---|---|
| Current (all warm) | ~$4,960 | ~$59,500 |
| With cold storage | ~$1,500-1,800 | ~$18,000-21,600 |
| Potential Savings | ~$3,200/month | ~$38,000/year |
What I Learned
- Data analysis with AI is powerful — Extracting and analyzing Excel data, even when tools couldn’t read the file directly, was possible by extracting XML from the xlsx archive
- Proactive cost modeling is key — Many organizations don’t calculate steady-state costs before implementing their backup strategy
- Quick wins exist — Cold storage tiering is a configuration change that can save 60-75% on long-term backups
- Visualization helps — Snapshot growth charts immediately reveal accumulation patterns
What’s Next
- Validate RPO/RTO requirements match backup retention
- Implement cold storage tiering on monthly/yearly backups
- Set up CloudWatch alarms for snapshot storage monitoring
Related Posts
Browser Automation Agents - Amazon Bedrock AgentCore
Enterprise workflows often require interacting with web applications that lack APIs. Traditional automation scripts are brittle and break when UIs change.
DevelopmentTime Series Forecasting App - Amazon Chronos-2
Building a production forecasting application without the complexity of traditional ML model training and feature engineering.
DevelopmentManaging Local Storage in the AI Development Era
How to identify, clean, and monitor the hidden storage consumers that come with AI-assisted development tools like Claude Code and Kiro
