NestJS Performance

Measure First

Optimize with evidence. Profile endpoints with load tools, inspect slow query logs, and track event-loop lag. Guessing often accelerates the wrong layer.

const app = await NestFactory.create(AppModule, new FastifyAdapter());
await app.listen(3000, '0.0.0.0');

Switching from Express to Fastify can improve throughput for many APIs, but measure on your workload.

High-Impact Wins

- Cache hot GET responses (CacheInterceptor / Redis)
- Avoid N+1 queries; select only needed columns
- Compress responses when payloads are large
- Use pagination everywhere lists can grow
- Prefer singleton providers; avoid unnecessary REQUEST scope

• REQUEST-scoped providers disable some Nest optimizations—use sparingly.
• DTO whitelisting reduces accidental large payloads.
• Enable HTTP keep-alive and proper connection pooling for databases.
• Offload CPU-heavy work to queues/workers.

Performance Cheatsheet

Common Nest performance levers.

Caching Strategically

Cache expensive, read-heavy, mostly-immutable responses. Invalidate on writes. Never cache personalized data under a shared key.

Serialization Cost

ClassSerializerInterceptor and large nested graphs can add CPU cost. Return lean DTOs for list endpoints.

Common Mistakes

• Micro-optimizing code before fixing missing DB indexes.
• Caching user-specific responses with a global key.
• Overusing request-scoped providers.
• Loading entire tables without pagination.

Key Takeaways

• Measure before optimizing Nest performance.
• Fastify, caching, and query efficiency are high-impact levers.
• Avoid unnecessary request scoping.
• Paginate and slim payloads by default.