Cost Optimization Audit

Step 1: Gather context

What workload or AWS environment would you like me to audit for cost optimization? Please share:

• Architecture overview (services used, rough monthly spend if known)
• Traffic patterns (steady, spiky, predictable growth, seasonal)
• Commitment status (Savings Plans, Reserved Instances, existing contracts)
• Budget constraints or targets (optional)

Step 2: Identify waste

• 🔴 High — significant ongoing waste (>20% of category spend or >$500/month)
• 🟡 Medium — moderate waste (5-20% of category spend or $100-500/month)
• 🟢 Low — minor waste or optimization opportunity

• Idle resources — unattached EBS volumes, unused Elastic IPs, idle load balancers, stopped instances with attached storage
• Over-provisioned resources — instances with <20% CPU/memory utilization, over-sized RDS instances, over-provisioned DynamoDB capacity
• Orphaned resources — old snapshots, unused AMIs, stale log groups, abandoned S3 buckets
• Redundant data transfer — cross-AZ traffic that could be avoided, NAT Gateway costs for S3/DynamoDB (use VPC endpoints)

Step 3: Evaluate pricing models

• Are Savings Plans or Reserved Instances used for steady-state workloads?
• Could Spot Instances cover fault-tolerant or batch workloads?
• Would serverless (Lambda, Fargate, Aurora Serverless) reduce cost for variable workloads?
• Are S3 storage classes optimized? (Intelligent-Tiering, Glacier for archives)
• Is there opportunity for Graviton-based instances? (better price-performance)
• Are data transfer costs optimized? (VPC endpoints, CloudFront, regional placement)

Step 4: Assess architecture efficiency

• Could caching reduce compute/database load? (ElastiCache, CloudFront, DAX)
• Are there synchronous calls that could be async? (SQS, EventBridge — reduces over-provisioning)
• Is data lifecycle managed? (S3 lifecycle policies, RDS snapshot retention, log expiration)
• Are environments right-sized for their purpose? (dev/test smaller than prod, scheduled scaling)
• Can non-production environments be scheduled off during nights/weekends?

Step 5: Quantify savings

• Current cost (monthly)
• Projected cost after change
• Savings ($ and %)
• Effort to implement (Low / Medium / High)
• Risk of the change
• AWS Service to use

Step 6: Produce the report

# Cost Optimization Audit: {Workload Name}

## Summary
- **Date**: {date}
- **Estimated current monthly spend**: ${X}
- **Potential monthly savings**: ${Y} ({Z}%)
- **Findings**: {A} High, {B} Medium, {C} Low

## Cost Optimization Scorecard
| Domain | Score (1-5) | Key Gap |
|--------|-------------|---------|
| Waste Elimination | {score} | {gap} |
| Pricing Model | {score} | {gap} |
| Architecture Efficiency | {score} | {gap} |
| Data Lifecycle | {score} | {gap} |
| Environment Management | {score} | {gap} |

## High-Impact Findings
{Each: what's wasteful, severity, current cost, projected savings, AWS service, effort}

## Quick Wins (< 1 week)
{Each: what to do, savings estimate, how to implement, AWS service}

## Foundation Improvements (1-4 weeks)
{Each: what to change, savings estimate, trade-offs, AWS service}

## Strategic Changes (1-3 months)
{Each: what to redesign, savings estimate, effort, risk, AWS service}

## Savings Summary
| Category | Current $/mo | Optimized $/mo | Savings $/mo | AWS Service |
|----------|-------------|----------------|--------------|-------------|
| Compute | {current} | {optimized} | {savings} | Compute Optimizer, Savings Plans |
| Storage | {current} | {optimized} | {savings} | S3 Intelligent-Tiering, Glacier |
| Data Transfer | {current} | {optimized} | {savings} | VPC Endpoints, CloudFront |
| Database | {current} | {optimized} | {savings} | RDS Reserved, Aurora Serverless |
| Other | {current} | {optimized} | {savings} | {relevant services} |
| **Total** | **{total}** | **{total}** | **{total}** | |

## Implementation Roadmap
| Priority | Action | Savings/mo | Effort | Risk |
|----------|--------|-----------|--------|------|
| 1 | {action} | ${savings} | Low | Low |
| 2 | {action} | ${savings} | Medium | Low |
| ... | ... | ... | ... | ... |

## Next Steps
{Concrete actions to start saving immediately}

Step 7: Offer follow-up

Would you like me to:

• Create a detailed right-sizing plan for a specific service?
• Model Savings Plans vs Reserved Instances for your usage pattern?
• Design a FinOps tagging and cost allocation strategy?
• Build a scheduled scaling policy for non-production environments?
• Estimate costs for an architectural alternative (serverless, containers, etc.)?