Availability and High Availability
QNSP is designed for high availability across all tiers.
Availability and High Availability
QNSP is designed for high availability across all tiers.
Availability targets
| Component | Target | Notes |
|---|---|---|
| Edge gateway | 99.99% | Multi-AZ, health-based routing |
| Auth service | 99.9% | Stateless, horizontally scaled |
| KMS service | 99.9% | HSM-backed, cached operations |
| Storage service | 99.9% | Replicated storage backend |
HA architecture
Stateless services
- Multiple replicas per service
- Load balancer distributes traffic
- Any replica can handle any request
Stateful components
- Database: Multi-AZ with automatic failover
- HSM: Clustered with replication
- Cache: Redis cluster mode
Health checks
All services expose health endpoints through edge-gateway:
GET /proxy/<service>/health— basic livenessGET /edge/<service>/health— alternative health path
Health check endpoints (GET/HEAD) bypass bot protection and rate limiting for reliable monitoring.
Failover
- Automatic failover within region
- Cross-region failover requires
failover-regionadd-on - RTO: < 5 minutes (within region)
- RPO: 0 (synchronous replication)