Transparency of Costs — Governance on Harmony
Last updated: January 23, 2026
Public document for community presentation
Stack: Aragon OSX fork
Executive Summary
Performance Cost (Recommended)
| Category | Monthly Cost | Annual Cost |
|---|---|---|
| Compute Engine | $280 | $3.360 |
| Cloud Memorystore (Redis) | $35 | $420 |
| Cloud Pub/Sub (Message Queue) | $25 | $300 |
| Cloud Storage & Backup | $15 | $180 |
| Cloud DNS | $1 | $12 |
| Cloud Load Balancer | $18 | $216 |
| Cloud Monitoring & Logs | $20 | $240 |
| Sentry (Optional) | $29 | $348 |
| TOTAL PERFORMANCE (with Sentry) | $423/mo | $5.076/yr |
| TOTAL (without Sentry) | $394/mo | $4.728/yr |
Minimum Viable Cost (Startup)
| Category | Monthly Cost | Annual Cost |
|---|---|---|
| Compute Engine (2vCPU, 12GB RAM, 50GB SSD) | $110 | $1.320 |
| Cloud Memorystore (768MB) | $12 | $144 |
| Cloud Pub/Sub (Basic) | $5 | $60 |
| Cloud Storage | $5 | $60 |
| Cloud DNS | $1 | $12 |
| Cloud Load Balancer | $0 | $0 |
| Cloud Monitoring (Free) | $0 | $0 |
| Sentry (Optional) | $29 | $348 |
| TOTAL MINIMUM (with Sentry) | $162/mo | $1.944/yr |
| TOTAL (without Sentry) | $133/mo | $1.596/yr |
Detailed Comparison
| Component | Minimum | Performance | Difference |
|---|---|---|---|
| Scalability | Medium (1.000 req/min) | High (5.000+ req/min) | 5x |
| API Latency | 100-300ms | 50-100ms | 2-3x better |
| Uptime SLA | 95% | 99.9% | Enterprise-grade |
| Concurrent Users | ~500 | ~1.000 | 2x capacity |
| vCPU / RAM | 2vCPU / 12GB | 4vCPU / 16GB | Balanced vs Optimized |
| Storage | 50GB SSD | 100GB SSD | 2x more space |
| Backup | Manual | Automatic 24h | Full redundancy |
| Monitoring | Basic | Enterprise | Advanced alerts |
MINIMUM COST (Startup/MVP)
Compute Engine — $110/mo
Configuration:
Machine: n2-standard-2 (2 vCPU, 12GB RAM)
Region: us-central1 (Iowa)
Storage: 50GB SSD
OS: Ubuntu 22.04 LTS
Services: Node.js (API) + MongoDB (local) + Redis (local)
Commitment: Pay-as-you-go
Estimated: $110/mo
Capacity:
- ~1.000 requests/min at peak (up from 500)
- Single Node.js + Express server with multiple workers
- MongoDB running on same VM (sufficient for <50K DAOs)
- 2-3 worker processes
- Suitable for small to medium deployments
Advantages over previous minimum:
- More RAM → better MongoDB performance
- Better CPU allocation → handles concurrent requests better
- Larger storage → more data before needing cleanup
- Production-ready baseline
Limitations:
- No auto-scaling
- No load balancer
- No geographic redundancy
- Manual SSH deployments
Cloud Memorystore (Redis) — $12/mo
Configuration:
Plan: Basic (768MB)
Region: us-central1
Redis Version: 7.2
Replication: None
Persistence: Disabled
Estimated: $12/mo
Use Cases:
- API response caching (5min TTL)
- Simple rate limiting
- ~5K simultaneous keys
Cloud Pub/Sub — $5/mo
Configuration:
Topics: 3 (events, jobs, notifications)
Subscriptions: 5
Throughput: 10K msgs/day
Retention: 7 days
Estimated: $5/mo
Cloud Storage — $5/mo
Configuration:
Storage: ~2GB data
Transfer: 5GB/mo outbound
Class: Standard
Auto-backup: None
Estimated: $5/mo
Cloud DNS — $1/mo
DNS:
Public Zone: governance.one
Records: A, MX, NS, CNAME (~10)
Query Volume: 100K/mo
Estimated: $1/mo
Cloud Monitoring (Free)
Monitoring:
Custom Metrics: ~50
Logs: Up to 50GB/mo
Dashboards: 10+ custom
Estimated: $0/mo
Sentry (Optional) — $29/mo
Configuration:
Plan: Pro ($29/mo)
Events/mo: 500K
Retention: 90 days
Alerting: Yes
Estimated: $29/mo (OPTIONAL)
TOTAL MINIMUM: $133-162/mo (~$1.596-1.944/yr)
Best for: MVP, testnet, production-grade startup, 500-5K users
Ready for: Early adopters, beta testing, initial DAO deployments
PERFORMANCE COST (Production)
Compute Engine — $280/mo
Configuration:
Machine Type: n2-standard-4 (4 vCPU, 16GB RAM) × 2
Regions: us-central1 (primary) + europe-west1 (backup)
Storage: 100GB SSD each
Managed Instance Group: Auto-scaling (2-5 instances)
Load Balancer: Global
Services: Node.js (API) + MongoDB (local) + Redis (local)
OS: Ubuntu 22.04 LTS with PM2
Estimated: $140/mo (us) + $140/mo (eu) = $280/mo
Capacity:
- ~5.000 requests/min at peak
- 2 workers per machine (4 total)
- Auto-scaling 2-5 instances on demand
- MongoDB replicated across instances
- <100ms latency in 99% of cases
Features:
- Cloud CDN integrated
- Automatic health checks
- Multi-region support
Cloud Memorystore (Redis) — $35/mo
Configuration:
Plan: Standard (5GB)
Regions: us-central1 + read-only replica (eu)
Redis Version: 7.2
Replication: YES (HA)
Persistence: RDB 1h + AOF
Auto-failover: Yes
Estimated: $35/mo
Use Cases:
- API response caching (30min TTL)
- Distributed rate limiting
- Session management
- ~50K simultaneous keys
Cloud Pub/Sub — $25/mo
Configuration:
Topics: 10 (segmented by type)
Subscriptions: 20+ (with dead-letter queues)
Throughput: 500K msgs/day
Retention: 24 hours
Dead Letter Queue: Yes
Estimated: $25/mo
Cloud Storage (Backup & Logs) — $15/mo
Configuration:
Storage: ~10GB data + backups
Transfer: 100GB/mo outbound
Class: Standard (hot data)
Replication: Multi-region
Versioning: Enabled
Estimated: $15/mo
Cloud DNS — $1/mo
DNS:
Public Zones: 3 (governance.one, .country, .xyz)
Records: ~30 (A, MX, NS, CAA, TXT)
DNSSEC: Enabled
Query Volume: 10M/mo
Estimated: $1/mo
Cloud Load Balancer — $18/mo
Configuration:
Type: Global HTTP(S) Load Balancer
Forwarding Rules: 1
Backend Services: 3 (API, WebSocket, Admin)
Health Checks: Continuous
CDN: Cloud CDN integrated
DDoS Protection: Basic
Estimated: $18/mo
Cloud Monitoring & Logging — $20/mo
Configuration:
Uptime Checks: 5 endpoints
Custom Metrics: 100+
Log Retention: 30 days
Alerting Policies: 20+
Dashboards: 5+ custom
Error Budgets: Enabled
Estimated: $20/mo
Sentry (Optional) — $29/mo
Configuration:
Plan: Pro ($29/mo)
Events/mo: 500K
Retention: 90 days
Team Members: Unlimited
Alerting: Webhooks, Slack, Email
Estimated: $29/mo (OPTIONAL)
TOTAL PERFORMANCE: $394-423/mo (~$4.728-5.076/yr)
Best for: Production-grade, 5K-100K users, 99.9% SLA
Features: Auto-scaling, redundancy, 24/7 backups, advanced alerts
Scaling Roadmap
Months 1-3 (MVP):
├─ Setup: Minimum Cost ($133/mo)
├─ Users: <500
├─ MongoDB: Local on VM (12GB RAM sufficient)
└─ Focus: Product validation
Months 4-6 (Growth):
├─ Upgrade: ~$200-250/mo (add second instance + HA)
├─ Users: 500-5K
├─ MongoDB: Replication across 2 VMs
└─ Focus: Stability
Months 7+ (Scale):
├─ Full Performance: $394/mo
├─ Users: 5K-100K
├─ MongoDB: 3+ replicas distributed
└─ Focus: Global redundancy
Optimization Strategies (Implemented)
MongoDB local: Eliminates Cloud SQL cost (~$150/mo saved)
Native Harmony RPC: Zero RPC provider cost (Alchemy/QuickNode eliminated)
Aggressive caching: 30min TTL saves ~40% database queries
Pub/Sub for async: Reduces main API load
Cloud Storage for backups: Automatic and cheap (<$15/mo)
Managed DNS: Simple and economical ($1/mo)
Right-sized minimum VM: 12GB RAM → better MongoDB performance out of the box
Proposed Stack (Google Cloud)
Infrastructure:
├─ Compute Engine (Node.js + MongoDB): $110-280/mo
├─ Cloud Memorystore (Redis): $12-35/mo
├─ Cloud Pub/Sub (Message queue): $5-25/mo
├─ Cloud Storage (Backups): $5-15/mo
├─ Cloud DNS: $1/mo
├─ Cloud Load Balancer: $0-18/mo
└─ Cloud Monitoring: $0-20/mo
Observability (Optional):
└─ Sentry: $29/mo
Harmony One:
├─ RPC: Public API (http://api.harmony.one)
├─ Indexing: Self-hosted
└─ Cost: $0
Total:
├─ MINIMUM: $133/mo (without Sentry)
├─ MINIMUM with Sentry: $162/mo
├─ PERFORMANCE: $394/mo (without Sentry)
├─ PERFORMANCE with Sentry: $423/mo
└─ ANNUALLY: $1.596 - $5.076
Performance Metrics by Tier
| Metric | Minimum | Performance |
|---|---|---|
| Requests/min (peak) | 1.000 | 5.000+ |
| Latency p95 | 100-300ms | 50-100ms |
| Uptime SLA | 95% | 99.9% |
| Concurrent Users | ~500 | ~1.000 |
| DAO Support | ~50K | 100K+ |
| Indexing Lag | 30-60s | 5-10s |
| Backup Frequency | Manual | Automatic 1h |
| Geo-redundancy | None | Yes (2+ regions) |
| Auto-scaling | None | Yes (2-5 nodes) |
| MongoDB Replication | None | Yes (3+ replicas) |
Hidden Costs to Avoid
❌ Data Transfer between regions: $0.12/GB (use same region)
❌ Disk snapshots: $0.05/GB/mo (clean up old snapshots)
❌ Static IP addresses: $3.50/mo each (use ephemeral)
❌ Cloud KMS encryption: $6/mo per key (use free tier)
❌ Custom machine types: 30% more expensive (use predefined)
❌ Egress to internet: $0.12/GB (minimize transfer)
Recommended Optimizations
Implement from day one:
- GCP Committed Use Discounts: Save 25% on Compute (1-year contract)
- Cloud Storage class downgrade: 30 days → Archive (90% savings after 30d)
- MongoDB index optimization: Reduce query time 50-80%
- Query result caching: 5-30min TTL (saves 80% database load)
- Log aggregation: Eliminate redundant logs (save $5-10/mo)
- Unused resource alerts: Auto-disable unused resources
- Instance schedules: Shutdown during off-hours (save 20% on dev)
12-Month Projection
Scenario 1: MVP (Minimum Cost - without Sentry)
├─ Months 1-3: $133 × 3 = $399
├─ Months 4-6: $200 × 3 = $600 (upgrade)
├─ Months 7-12: $350 × 6 = $2.100 (transition)
└─ TOTAL: ~$3.099 (Year 1)
Scenario 2: Production from day one (Performance - without Sentry)
├─ Months 1-12: $394 × 12
└─ TOTAL: ~$4.728 (Year 1)
Scenario 3: Production with Sentry from day one
├─ Months 1-12: $423 × 12
└─ TOTAL: ~$5.076 (Year 1)
Cost savings vs. Ethereum mainnet deployment:
├─ RPC costs eliminated: $30-50/mo (no paid RPC needed)
├─ Gas costs eliminated: ~$200-500/mo (Harmony gas is free)
└─ TOTAL ANNUAL SAVINGS: ~$2.760-6.600/yr
Comparison with Alternatives
| Solution | Monthly Cost | Scalability | Decentralization |
|---|---|---|---|
| Aragon OSx on Harmony (Current) | $394-423 | Excellent | Medium |
| Ethereum L1 | $900+/mo | Limited | Medium |
| Fully Decentralized (The Graph) | $2.000+ | Medium | Excellent |
| Polygon/zkSync | $500-800/mo | Great | Medium |
| Centralized Cloud (AWS) | $600+/mo | Excellent | Low |
Conclusion: Harmony offers the best cost-benefit for decentralized DAO governance infrastructure.
Funding Model
Who Covers the Costs?
This infrastructure is maintained by the Harmony ecosystem, option available:
- RMC Board (Recovery DAO): Primary funding source for operational costs
- Harmony Core: Can support critical infrastructure components
- Community contribution: Voluntary donations/sponsorships welcome but not required
Cost Breakdown by Stakeholder
Monthly Operating Budget: $394/mo (Performance tier)
├─ Infrastructure (primary requirement):
│ ├─ Compute + Storage: $311/mo (79%)
│ └─ Monitoring + Backup: $83/mo (21%)
│
└─ Observability (optional enhancement):
└─ Sentry error tracking: $29/mo
Free services available to all DAOs:
- DAO creation and management on Aragon OSx
- Governance proposals and voting
- Plugins (default) ecosystem access
- Community support and documentation
Scaling Indicators
Upgrade to Performance when:
├─ API latency consistently > 300ms
├─ Error rate > 1% during peak hours
├─ Concurrent connections > 1.000
├─ Database queries > 100/second
├─ Disk usage > 70% of capacity
└─ More than 5K unique users/mo
Locally Installable Components
These components run inside the VM with no additional costs:
On Compute Engine:
├─ MongoDB 7.0 (replica set for HA)
├─ Redis 7.2 (in-memory, disk-backed)
├─ Node.js 20 LTS (API + workers)
├─ PM2 (process manager)
├─ Nginx (reverse proxy)
├─ Prometheus (local metrics)
├─ Docker (builder env)
└─ ELK Stack (logs - optional, local)
Benefit: Full control, no vendor lock-in, maximum performance.
Support Infrastructure
This infrastructure enables Aragon OSx governance on Harmony at minimal cost globally. It’s maintained transparently for the entire Harmony ecosystem.
Questions? Open an issue or reach out on telegram.
Transparent costs = Sustainable community infrastructure.