| name | capacity-report |
| description | Generate a monthly capacity and cost report for a service or AWS account, with rightsizing recommendations |
Generate a capacity and cost report. Use the cost-analysis MCP tool and cloudformation MCP tool where available.
If a service name or stack is not in context, ask: "Which service or AWS account should I report on? Provide the stack name, service name, or account ID."
Report scope
Ask if not clear:
- Service report: one ECS service, Lambda function, or RDS cluster
- Account report: all resources in an AWS account for the month
Data to collect (use MCP tools where available)
- Cost breakdown — last 30 days, grouped by service and tag
- Compute utilisation — ECS CPU/memory %, EC2 CPU %, Lambda invocations/duration
- Database utilisation — RDS CPU, connections, storage used vs provisioned
- Auto-scaling activity — scale-out and scale-in events, min/max/average task count
- Data transfer — outbound internet, cross-AZ, NAT Gateway
Output format
Cost Summary — [Month]
| Category | This month | Last month | Change |
|---|
| Compute (ECS/EC2) | $XXX | $XXX | +/-X% |
| Database (RDS) | $XXX | $XXX | |
| Data transfer | $XXX | $XXX | |
| Storage (S3/EBS) | $XXX | $XXX | |
| Other | $XXX | $XXX | |
| Total | $XXX | $XXX | +/-X% |
Capacity Utilisation
| Resource | Avg utilisation | Peak utilisation | Recommendation |
|---|
| ECS service (CPU) | 42% | 78% | Appropriate — no action |
| ECS service (Memory) | 85% | 95% | Consider increasing memory |
| RDS (CPU) | 12% | 28% | Potential downsize opportunity |
| RDS (Storage) | 45% | 45% | On track — ~8 months until full |
Rightsizing Recommendations
For any resource with average utilisation <30% (potential downsize) or >70% sustained (potential upsize):
| Resource | Current size | Recommended size | Monthly saving | Action |
|---|
| RDS db.r6g.large | db.r6g.large | db.r6g.medium | ~$85/month | Low risk — schedule maintenance window |
| ECS Memory | 2048 MB | 3072 MB | -$12/month | Increase to prevent OOM events |
Cost Anomalies
Flag any day-over-day cost increases >20% or unusual spend patterns:
- "Data transfer cost spiked on [date] — coincides with [deployment/event]"
- "NAT Gateway cost increased 3x — investigate traffic pattern change"
Top 5 Optimisation Actions
Ranked by estimated monthly saving:
- [Action] — saves ~$XXX/month — [effort: Low/Medium/High]
- ...
Forecast
Based on current trends: next month estimated cost $XXX (+/-X% vs this month).
Key assumptions: [list any seasonal or event-driven factors]
