Official performance targets and measurements for nself.

Last Updated: January 31, 2026 Version: 0.9.8

---

Operation	Target	Measured	Status
Build (clean, no services)	< 10s	7s	✅
Build (clean, all services)	< 30s	25s	✅
Build (incremental)	< 5s	3s	✅
Config generation only	< 2s	1s	✅

Operation	Target	Measured	Status
Start (required services only)	< 30s	25s	✅
Start (all services)	< 60s	50s	✅
Stop (all services)	< 10s	6s	✅
Restart (single service)	< 5s	3s	✅
Status check	< 1s	0.5s	✅

Operation	Target	Measured	Status
Migration (up, 10 files)	< 5s	3s	✅
Migration (down, 10 files)	< 5s	4s	✅
Seed (1000 rows)	< 10s	7s	✅
Backup (1GB database)	< 60s	45s	✅
Restore (1GB backup)	< 90s	70s	✅

Endpoint	Target	Measured (p50)	Measured (p95)	Status
GraphQL (simple query)	< 50ms	20ms	30ms	✅
GraphQL (complex query)	< 200ms	100ms	150ms	✅
GraphQL (mutation)	< 100ms	50ms	80ms	✅
Auth login	< 100ms	60ms	80ms	✅
Auth verify token	< 50ms	25ms	35ms	✅

Scenario	Target	Measured	Status
Concurrent users	1000	1200	✅
Requests/second	5000	6000	✅
Error rate	< 0.1%	0.05%	✅
95th percentile latency	< 200ms	150ms	✅

---

• CPU: 4 vCPU
• RAM: 8GB
• Disk: SSD
• Network: 1Gbps

• All optional services enabled
• Monitoring bundle enabled
• 2 custom services
• Production mode

• Duration: 5 minutes
• Ramp-up: 1 minute
• Think time: 1-3 seconds
• Distribution: Normal

---

Version	Clean Build	Incremental
v0.9.6	32s	5s
v0.9.7	28s	4s
v0.9.8	25s	3s

Version	Simple Query	Complex Query
v0.9.6	45ms	180ms
v0.9.7	35ms	160ms
v0.9.8	30ms	150ms

---

• Use incremental builds when possible
• Disable unused services
• Use build caching (v0.9.8+)

• Configure PgBouncer for connection pooling
• Enable Redis caching
• Use CDN for static files
• Optimize database queries

• Create appropriate indexes
• Use prepared statements
• Enable query caching
• Regular VACUUM ANALYZE

---

See Also:

• Quality Metrics
• Testing Strategy