Skip to content

COST.md

Latest commit

 

History

History
148 lines (114 loc) · 5.1 KB

COST.md

File metadata and controls

148 lines (114 loc) · 5.1 KB

CodeRipple Cost Analysis

Comprehensive cost breakdown for the CodeRipple serverless analysis pipeline.

AWS Lambda Costs

Free Tier Benefits

AWS Lambda provides generous free tier allowances:

  • 1 million requests per month (forever)
  • 400,000 GB-seconds of compute time per month (forever)

Hermes Lambda (Event Logging)

  • Memory: 256 MB
  • Typical execution: 1-2 seconds per event
  • Usage estimate: 1,000-10,000 events/month
  • Cost: FREE (well within free tier limits)

Receptionist Lambda (Repository Processing)

  • Memory: 1GB (1,024 MB) - Updated from 512MB for better performance
  • Ephemeral Storage: 5GB (5,120 MB) - Testing configuration for 1 week
  • Typical execution: 30-120 seconds per repository
  • Usage estimate: 100-1,000 repositories/month

Cost Analysis:

  • Moderate usage (1,000 invocations/month, 60 seconds each):
    • GB-seconds: 60,000
    • Cost: FREE (within 400K GB-seconds free tier)
  • Heavy usage (10,000 invocations/month, 120 seconds each):
    • GB-seconds: 1,200,000
    • Free: 400,000 GB-seconds
    • Paid: 800,000 GB-seconds × $0.0000166667 = $13.33/month

Memory Upgrade Cost Impact

Upgrading Receptionist from 512MB to 1GB:

  • Performance benefit: 2x CPU allocation, faster git operations
  • Cost impact: Doubles compute cost but faster execution likely reduces total GB-seconds
  • Net effect: Minimal cost increase, significant performance gain

Lambda Ephemeral Storage Costs

Pricing Structure

  • First 512 MB: Included in Lambda pricing (no additional cost)
  • Additional storage: $0.0000000309 per GB-second

Receptionist 5GB Storage (Testing Phase)

  • Additional storage: 5GB - 0.5GB = 4.5GB
  • Cost per second: 4.5GB × $0.0000000309 = $0.000000139 per second

Weekly cost examples:

  • Light usage (100 invocations, 30 seconds each): ~$0.0004
  • Moderate usage (1,000 invocations, 60 seconds each): ~$0.008
  • Heavy usage (10,000 invocations, 120 seconds each): ~$0.17

Total for 1-week testing period: $0.01-0.20 (negligible)

Post-testing optimization: Reduce to 1-2GB based on actual usage patterns to minimize ongoing costs.

S3 Storage Costs

Cabinet S3 Bucket (Public Website)

  • Storage class: Standard
  • Usage: Event logs, static website files
  • Size estimate: <1GB for event logs
  • Cost: ~$0.02/month (minimal)

Library S3 Bucket (Repository Storage)

  • Storage class: Standard (with lifecycle policies planned)
  • Usage: Cloned repositories, analysis files
  • Size estimate: Varies by repository count and size
  • Cost factors:
    • Storage: $0.023 per GB/month
    • PUT requests: $0.0005 per 1,000 requests
    • GET requests: $0.0004 per 1,000 requests

Estimated monthly costs:

  • Light usage (10GB storage): ~$0.25/month
  • Moderate usage (100GB storage): ~$2.50/month
  • Heavy usage (1TB storage): ~$25/month

EventBridge Costs

Custom Events

  • Pricing: $1.00 per million events
  • Usage estimate: 1,000-10,000 events/month
  • Cost: FREE (well under 1 million events)

API Gateway Costs

REST API (Gatekeeper)

  • Pricing: $3.50 per million API calls
  • Usage estimate: 100-1,000 webhook calls/month
  • Cost: FREE (minimal usage)

Total Monthly Cost Estimates

Development/Testing Phase

  • Lambda: FREE (within free tier)
  • S3: $0.25-2.50 (depending on repository storage)
  • EventBridge: FREE
  • API Gateway: FREE
  • Total: $0.25-2.50/month

Production Phase (Heavy Usage)

  • Lambda: $13.33 (Receptionist only, others free)
  • S3: $25 (1TB repository storage)
  • EventBridge: FREE
  • API Gateway: FREE
  • Total: ~$38/month

Cost Optimization Strategies

Immediate (Week 1)

  1. Monitor ephemeral storage usage - Reduce from 5GB to optimal size
  2. Track repository sizes - Understand actual storage needs
  3. Monitor Lambda execution times - Optimize memory allocation

Short-term (Month 1)

  1. Implement S3 lifecycle policies - Move old repositories to cheaper storage
  2. Optimize Lambda memory - Right-size based on performance metrics
  3. Repository cleanup - Delete processed repositories after analysis

Long-term (Month 3+)

  1. S3 Intelligent Tiering - Automatic cost optimization
  2. Lambda Provisioned Concurrency - If consistent traffic develops
  3. Reserved Capacity - For predictable workloads

Cost Monitoring

CloudWatch Metrics to Track

  • Lambda execution duration and memory usage
  • S3 storage utilization and request patterns
  • EventBridge event volume
  • API Gateway request counts

Alerts to Set Up

  • Monthly spend exceeding $50
  • S3 storage growth exceeding 500GB
  • Lambda execution errors or timeouts

Conclusion

CodeRipple is designed to be extremely cost-effective:

  • Development phase: Under $3/month
  • Production phase: Under $40/month even with heavy usage
  • Free tier benefits: Significant cost savings for moderate usage
  • Serverless architecture: Pay only for actual usage, no idle costs

The serverless approach ensures costs scale with usage while maintaining excellent performance and reliability.